How the Internet Is Disrupting Metanarratives and Human Society

This article, first published in 2020, originates from a Chinese perspective and offers a philosophical critique of the internet. The title of the original article is “Is the Internet a detour in human history?
Upon its release, it quickly garnered over 500,000 reads in the Chinese-speaking world, sparking interest among Chinese internet researchers.
Its critique begins with the public’s view of the Chinese internet industry, yet its insights are not confined to China alone. Instead, it explores how the internet as a technology itself fractures human society.
This English version has been translated by ChatGPT and may contain errors or inaccuracies.

Democratic countries believe that the Internet destroys democracy, and authoritarian countries believe that the Internet destroys dictatorships.

How could this be possible?

1. The Darkest Hour

1.1 Is the Internet an Authoritarian Technology?

In April 2019, The New York Times published an article titled “The Only Answer Is Less Internet,” marking a new peak in the critical stance that Western mainstream media held towards the internet industry.

In the span of just April 2019, around 20 op-eds and critiques—not including news articles—emerged from Western mainstream outlets such as The New York Times, The Washington Post, VICE, Buzzfeed News, NBC, Harvard Business Review, Fortune, and The Verge, to name a few, all presenting a negative view of the broader internet sector.

Indeed, this wave of critical discourse has been ongoing for about a year since the Facebook-Cambridge Analytica scandal broke, challenging the previous narrative.

Contrary to what most people who don’t directly engage with foreign media might think, the internet industry has long since shed its early Western image as a beacon of freedom and democracy, morphing instead into an almost mythical source of all evils.

The article in The New York Times discusses two perceived development paths for the internet: one, a Western model, supposedly driven by the “invisible hand” of the free market, and the other, an Eastern model characterized by stringent regulation, strict market entry requirements, and heavy-handed management.

However, the piece suggests that this dichotomy is flawed. Both the Western and Eastern internet landscapes seem to have converged on a similar endpoint: centralization of power, reducing citizens to mere consumers, infringing on privacy, and manipulating public opinion.

From the author’s perspective, the evolution of the Western internet into its current state without a single government, corporation, or individual driving it is more alarming than the situation with the Eastern internet. This suggests that the issues associated with the internet—centralization, privacy invasion, consumer reductionism, and opinion manipulation—are not anomalies but inevitabilities for societies that have embraced this technology.

This notion starkly contrasts with the earlier ideals of equality, freedom, and democracy that were projected onto the internet. Even more fundamentally, within the Chinese context, there’s a prevalent belief in the “neutrality of technology,” meaning technology itself should not carry any inherent values or ideologies.

While I don’t fully agree with all the logic presented in “The Only Answer Is Less Internet,” it serves as a provocative starting point for our discussion.

Let’s begin with a tangible example: the internet and privacy.

In 2018, Robin Li, CEO of Baidu, unintentionally sparked a still unresolved debate with his comment about trading privacy for convenience at the China Development Forum. Contrary to the optimistic belief that “technology is inherently good but corrupted by commercial enterprises,” this stance is increasingly untenable.

Today, the infringement of user privacy by internet products and companies penetrates both their business models and user experiences, making public demands for privacy both unrealistic and impossible to fulfill.

I. The Business Impossibility

Since its inception, the internet has predominantly presented itself as a free-to-use platform, where most users can access services without paying a fee. However, “free” is not a business model; advertising is. In recent years, some users, lacking an understanding of business logic, have come to believe:

“I’ve contributed to the popularity of a company’s product by using it; therefore, it’s unfair for them to use my data for advertising.”

But the primary commercial value of this “popularity” lies in its advertising potential, where the competitive edge of internet advertising is precisely in leveraging user privacy.

Although internet advertising existed before the advent of big data technology, it’s a fact that mainstream internet companies only began turning significant profits when they started to “invade” user privacy for targeted advertising. Before the era of user profiling, personalized recommendations, and precision targeting, internet ads were much less effective, offering minimal value to advertisers.

Thus, the choice becomes clear: users must pay with their privacy to receive services, or without this trade-off, companies fail and services shut down.

An idealized alternative would be to transform services that rely on advertising revenue through user data into directly paid services. In essence, based on the costs of operating websites and apps, as well as development needs, companies would set a price to charge users directly, similar to the current model for cloud storage services.

However, this model doesn’t apply universally, such as with search engines, and for some services, it could lead to prohibitively high direct costs for users.

A straightforward calculation can illustrate the impracticality of the no-ad, user-pays model. Take Baidu as an example: in the fourth quarter of 2018, its “online marketing revenue,” which primarily comes from advertising, was 21.2 billion yuan. According to third-party statistics, during the same period, the monthly active users of the Baidu App were about 300 million.

This means if Baidu were to offer a version of its app without ads and without collecting user data, it would need to charge users at least 70 yuan per quarter to maintain its current level of revenue.

This is just the cost for one product from one of the BAT companies (Baidu, Alibaba, Tencent). If you truly desire a completely “clean” internet experience free from privacy concerns, the likelihood is that you simply couldn’t afford the costs required to sustain such products.

More likely than not, users would not opt for these paid products.

The internet is inherently competitive. As long as there is at least one product in the market using investor funds to subsidize a free user experience without adding ads, other products that start charging fees are unlikely to retain their users.

II. The Impracticality on the Product Level

While there might seem to be feasible solutions to the business impracticalities, achieving the goal of not infringing on privacy is even more challenging on a technical level.

Since 2015, when Google DeepMind’s artificial intelligence, AlphaGo, defeated a human Go champion, the third wave of AI began to crest. However, the true starting point of this AI wave can be traced back to 2006, with the introduction of the concept of deep learning networks.

AlphaGo’s victory was more of a culmination than a beginning, signaling the end of the third AI wave. Since then, we’ve seen an influx of AI technologies based on deep learning being integrated into various products.

By the end of 2019, two key insights about this AI wave had emerged:

Firstly, the limitation—we can’t expect the emergence of the kind of strong AI found in science fiction from this wave of AI.

Secondly, the cost—if you want convenience, you have to give up data.

Since 2015, despite the proliferation of mobile internet leading to an exponential increase in the amount of information on the internet, the term “information explosion,” which was frequently mentioned in the early days of the internet, has seldom been heard.

Artificial intelligence algorithms have penetrated nearly every product we use, filtering and selecting vast amounts of content, services, goods, and even friends for us. They save us significant amounts of time and money, substantially increasing the efficiency of internet use.

Today, it’s virtually impossible to completely eliminate AI algorithms from internet products. Doing so would fundamentally dismantle many product functionalities:

  • Imagine a Taobao(Amazon) where layouts are manually arranged, and products are listed in the order they’re added;
  • Imagine a Meituan-Dianping(Yelp) that only shows food delivery options within a 1 km radius of your location, sorted solely by distance;
  • Imagine a news app that’s manually curated and updates its content only a few times a day on a set schedule;
  • Imagine a TikTok that randomly throws a video at you from a content pool, regardless of your preferences;
  • Imagine a Zhihu(Quora) that can only retrieve content through search queries (which would be exceedingly difficult);

Even if internet companies were willing to make such changes, the collective choices of the market or users would likely indicate that this path is not viable.

Therefore, while it’s understandable for regulatory bodies and the public to demand privacy, safety, and efficiency from the internet sector, such expectations are not reasonable. This situation is akin to the dilemma faced by mobile operators who are often told, “I don’t want a cell tower near me, but I still want signal,” a demand that is currently unfeasible both technologically and commercially.

The conflict between privacy and algorithms is just one of the macro challenges facing the internet. This brings us to a higher level question: Is “technology neutrality” a fallacy?

Indeed, with the current state of internet technology, it appears so.

It’s important to note that “technology neutrality” in the Chinese context differs from “net neutrality” in the English context, as well as the “technology neutrality principle” established by the 1984 Universal City Studios, Inc. v. Sony Corporation of America case.

It more closely resembles the Safe Harbor provision of the United States’ Digital Millennium Copyright Act (DMCA) of 1998, which essentially states that internet service providers cannot be held liable for copyright infringement by users if they are unaware of the infringing activities.

This concept has been extended beyond copyright to include web scraping, big data, finance, and other industries, often serving as a shield behind the notion that “ignorance is innocence.”

However, distinct from the well-known Safe Harbor provision, the “Red Flag” rule from the DMCA, which applies when a service provider is aware of infringing activities, holds more practical significance in reality.

The “Red Flag” principle stipulates that the safe harbor protection applies only if the internet service provider (ISP) is not aware, nor should be aware, that the information transmitted or linked through its service constitutes or contains pirated content.

For a practical example, consider the BitTorrent protocol. The company that invented the BitTorrent protocol could be considered truly “technologically neutral” because it merely created a technology without the ability to know what third parties or users are transmitting using this protocol.

However, in the BitTorrent ecosystem, there’s a need to establish Tracker servers. While Trackers don’t host any content themselves, they are essential for establishing transmission channels between users, thus enabling the potential to know the content being transmitted.

Even though the Tracker servers don’t contain pirated content and the pirated content transmitted between users isn’t provided by the owners of the Tracker servers, the fact that the owners “can know” about the pirated transmissions means they are not protected by the safe harbor principle if they do not actively regulate the content being distributed. Hollywood can, therefore, sue the owners of Tracker servers and win.

Thus, the technological neutrality of the BitTorrent protocol exists only in a vacuum and is lost once the technology is utilized. This aligns with the principle of technological neutrality established by the Sony Betamax case, where Sony was not held responsible for pirated videotapes because it had no control over them.

A technology can only be considered “neutral” and exempt from liability if its inventors or users are unable to control its societal impacts, whether positive or negative.

If technology is not neutral, does it carry ideology? Following this line of reasoning, it’s argued that technology does indeed embody ideology. This means that regardless of whether technology can be used for good or ill, and regardless of the intentions of its creators, a technology is inherently more suited for certain applications than others. Continuing with the BitTorrent protocol example, it’s observed that in practice, BitTorrent is more widely used for distributing pirated content than for legitimate business services, demonstrating its broader application in one area over another.

The notion that technology carries ideology is not a new analytical framework. Marshall McLuhan famously stated “the medium is the message” in his seminal work “Understanding Media” in 1964, suggesting that media inherently carries ideology. In the context of modern internet products, all products act as mediums.

To simplify, “the medium is the message” implies that the medium is not a neutral conduit. The way an idea or expression is conveyed through a medium determines its interaction with the audience, and this mode of interaction is in itself a form of content.

Taking a closer example from the media industry, “internet media” and “traditional media” represent two technologically distinct forms carrying different ideologies. The latter is limited by technology to one-way communication, while the former, enabling two-way information exchange, naturally encourages a shift from information receivers to content producers.

On a practical level, the main difference lies in how internet news content is not always the full representation of the ideas being communicated. The direction of the comments section, or even the allowance of comments, is considered part of the conveyed message.

Acknowledging the “ideology of the internet” leads to the realization that the spirit of the internet might not be as benevolent as initially imagined. The contradiction between efficiency and privacy highlighted by Robin Li is just one of the many irreconcilable contradictions of the internet.

Focusing on the drawbacks brought by the internet reveals that not all issues can be attributed to the subjective malevolence of internet companies. Worse, it suggests that some of the internet’s negative aspects are an inevitable outcome of its technological development.

This reflects Marx’s theory of the dialectical relationship between productive forces and production relations on a micro level, where the use of bronze and iron led to the transition from slavery to feudalism, and the invention of the printing press inevitably triggered the Enlightenment, laying the groundwork for the bourgeois revolution.

Regardless of the original intentions behind a technology or how its users wish to employ it, its operation inherently impacts society beyond merely improving production efficiency.

Understanding this is crucial to grasping the current situation faced by the internet and the remaining content of this discussion.

1.2 Why Critique the “Original Sin” of the Internet Now?

The internet is certainly not a technology imbued with authoritarian ideology, nor are internet companies a cohort of malevolent tech overlords driving such a narrative. However, this doesn’t imply everything is fine and the world is at peace.

The significance of the article “The Only Answer Is Less Internet” lies not in its literal call for humanity to reduce its reliance on the internet but rather in sparking a reflection on the pressing issues we face today.

What is this pressing issue? To understand it, we need to rewind a bit, back to the launch of the iPhone.

In January 2007, when Steve Jobs unveiled the first-generation iPhone at the Macworld convention in San Francisco, the level of excitement from the industry, media, and consumers surpassed that of any subsequent consumer electronics launch. Yet, no one could foresee that Apple had initiated the last technological marvel in consumer technology since the dissolution of the Soviet Union. In the decade following the iPhone’s introduction, no other product has so thoroughly transformed our lives, work, and even the way society operates as the smartphone has.

The mobile internet marked a second explosion of the internet, concluding this particular technological revolution by maximizing the transformative effects of the internet on human society. As mobile internet became universally accessible, the benefits the internet could offer society each subsequent day began to diminish rather than increase.

Hence, the “pressing issue” mentioned earlier is this: with the conclusion of this round of technological revolution, what do we do now?

The impact of each technological transformation on economic growth or societal change generally stems from two dimensions:

First, the capability for technological fission, or in layman’s terms, how many subsequent technologies are ignited by this technology. For instance, the mobile internet spurred the development of communication technology, display technology, cloud computing, big data, blockchain, artificial intelligence, and other related technologies. Each new secondary technological breakthrough adds fuel to the flames of technological revolution.

Second, the capacity for technological dissemination, meaning how many people are brought into the fold of the technological revolution. The lower the barriers to technology adoption, the more people can be involved, and the longer the duration of the technological revolution. For example, before the advent of the mobile internet (2008), there were fewer than 800 million internet users. Ten years after the birth of mobile internet in 2019, the number of internet users reached 4.38 billion.

Returning to the technological transformation of “Internet-Mobile Internet,” the peak of this chain reaction was reached around 2018-2019—everyone who could use the internet was using it, and every possible service that could be connected to the internet had been connected.

Throughout the burning process of this technological revolution, everything technology did was right because it visibly enhanced the welfare of society.

Despite early criticisms of Toutiao (a news aggregation app) as an “algorithm-driven trash heap,” it’s undeniable that before Toutiao, no news app had ever reached hundreds of millions of daily active users. The democratization of information it facilitated cannot be overlooked.

Similarly, while platforms like TikTok and mobile games are often criticized for “corrupting the younger generation,” humanity has never before created such easily accessible and affordable forms of entertainment. The democratization of entertainment (happiness) they brought about is also significant.

This situation parallels the debate on genetically modified (GM) foods, where some anti-GM advocates argue that GM foods should not be consumed by humans until decades-long studies have conclusively proven their absolute safety. However, the reality is that GM foods have provided a lifeline for many in impoverished regions globally, making the cessation of GM food production akin to a mass culling.

In terms of progress, the fundamental question of “whether or not” something exists must be addressed before we can rightfully discuss “how to improve” it.

Just as we only started focusing on the negative effects of sugar, oil, and salt after we could produce a meal of 498 calories for 17 yuan (a standard Big Mac), the digital “sugar, oil, and salt” brought by the internet follows the same pattern.

With both the degree of technological fission and the level of technological dissemination reaching their limits, the basic question of “whether or not” has been solved. Thus, we’ve gradually begun to pay attention to the legitimacy of these advancements.

Starting in 2018, a significant moral crisis emerged across the global internet industry, affecting companies from Facebook to Google, from Didi to Baidu, and from Toutiao to Ctrip, so much so that 2018 was jokingly referred to as the “Year of Internet Company Apologies.”

This moral crisis spanned different countries and regions, various sectors of the internet, and companies of different ages. Their only commonality was their identity as internet companies.

The deep-seated reason for this crisis is the extinguishing of the aforementioned two chains, leading to a slowdown in the rigid welfare growth that the internet industry could bring to society.

Put simply, once everyone has grown accustomed to the conveniences brought by the internet, it’s time to start paying attention to its negative impacts.

Indeed, on a macro level, governments and academic institutions recognized the systemic risks following the peak of internet benefits earlier than the general public.

Governments worldwide (regardless of their initial stance) began to intensify research and formulate regulations on the internet and data privacy: the European Union’s General Data Protection Regulation (GDPR) effective from 2018, the Digital Single Market Copyright Directive of 2018, the United States’ Clarifying Lawful Overseas Use of Data Act (also effective from 2018), China’s Cybersecurity Law of 2017, the UK’s Online Harms White Paper (which will advance legislative agenda), and Russia’s Sovereign Internet Law of 2019, among others.

Democratic countries feel that the internet undermines democracy, while authoritarian regimes believe it disrupts authoritarian control. This suggests that the internet has begun to reveal its detrimental effects on a more practical, micro-level, provoking hostility from structures of governance of any form.

While the legislative motives behind the documents listed above vary, they all aim to reduce the internet’s negative impacts by “making it less interconnected than before” in practical terms.

Alarmingly, these types of laws have received mainstream approval from governing bodies to the public in both traditional democracies and authoritarian states, facing little resistance even in the freest countries.

It’s worth mentioning that regulatory policies like GDPR have evidently failed to produce positive outcomes. According to a March 2019 report by the National Bureau of Economic Research (NBER), GDPR resulted in a reduction in financing per deal for emerging, young, and growing EU businesses by 27.1%, 31.4%, and 77.3% respectively, with new firms losing $900,000 in investments per week and mature firms losing $7.1 million per week.

Regulations like GDPR represent a quantifiable compliance cost for multinational giants like Facebook and Google, but pose insurmountable barriers for Europe’s fledgling local internet businesses.

This discussion does not aim to champion any specific management approach or explore the establishment of an effective internet governance system. Instead, it seeks to explain why the internet, once universally admired, has become somewhat of a “public enemy” in recent years.

Only by clearly identifying the problems of the Internet can we find real solutions. The media, the public, academia, and governments have yet to provide a reasonable explanation of the problems caused by the development of the Internet.

2. Unfulfilled Promises

2.1 The Dispute Between Social Determinism and Technological Determinism

On February 8, 1996, in response to the 1996 Telecommunications Act of the United States, John Perry Barlow, founder of the Electronic Frontier Foundation, published the “Declaration of the Independence of Cyberspace”. The first paragraph of this article, which was reprinted 40,000 times on the early Internet, reads as follows:

Governments of the Industrial World, you weary giants of flesh and steel, I come from Cyberspace, the new home of Mind. On behalf of the future, I ask you of the past to leave us alone. You are not welcome among us. You have no sovereignty where we gather.

A Declaration of the Independence of Cyberspace

This represents a classic technological determinism manifesto, embodying the early internet practitioners’ utopian dream about the internet. It envisioned cyberspace as a realm distinct from the physical world, one that should not interfere with the other.

This contradicts the objective law that any new technology will affect society and, in turn, be affected by it. As long as the devices that connect to cyberspace—computers, smartphones—and the infrastructure that supports it—servers, routers, physical cables—are produced by industrial machinery, no cyberspace can escape the governance of an “industrial government.” This logic is concisely but often overly simplified in the Chinese context with the phrase “the internet is not a lawless land.”

While the independence of cyberspace has not become a reality, it’s undeniable that the optimism driven by technological determinism has dominated the development of the internet.

Over the past 20 years, leaders and practitioners in the internet industry have basked in the glory of contributing societal welfare through their work. As mentioned earlier, before the breakdown of the two chain reactions, technological optimism was somewhat justified—as long as benefits outpaced the problems, the problems weren’t seen as significant.

However, when the two engines of growth objectively sputtered out, technological optimism and determinism became barriers to the internet industry’s self-correction and healing.

Just as acknowledging poverty is the first step towards solving it, the internet’s current dilemma is its foundation on a combination of technological determinism and optimism, led by those unwilling to recognize its darker technological aspects.

To put it simply, if you don’t even acknowledge the negative issues and impacts of the internet, there’s no way to address them.

Taking “technology for good” as an example, most people interpret this to mean “technology is inherently good,” suggesting that technological development naturally tends towards improving human society’s future. However, if your understanding stops there, you might be a classic technological optimist, possibly with a bent towards determinism.

This implies two underlying assumptions: technology inherently cannot do evil, and the goodness of technology will inevitably lead to societal goodness.

In reality, technology can be used for evil, and even technologies with no intention to harm can objectively contribute to negative societal impacts, such as nuclear energy.

Therefore, our advocacy should not only emphasize the natural role of technological development in promoting societal welfare but also prevent the application of technology in ways that reduce it, thereby mitigating the potential negative impacts of technological progress.

Technological optimists believe that even if technology introduces problems, the development of technology itself will further solve these issues. To some extent, this is true. Technology always drives human societal progress, but this usually refers to groundbreaking technological revolutions. Such revolutions don’t always cover the entire course of societal development; in reality, we must get accustomed to lengthy periods of stagnation between two groundbreaking technological shifts.


In recent years, numerous articles have plotted the timelines of humanity’s three technological revolutions on a single axis. By noting that the gap between the third and the second revolution is much shorter than between the second and the first, they infer a “rapidly accelerating geometric sequence” or “exponential curve.”

This optimism seems overly enthusiastic.

Brian Arthur, a pioneer of complexity economics, defines technology in his book “The Nature of Technology” as “a means to fulfill a human purpose through the capture and use of a phenomenon.” Accordingly, combination is considered a key source of technological progress. It’s like building something with LEGO blocks—we need wheels and transmissions to build a vehicle.

Technological optimists argue that the more building blocks we possess, the more likely we are to combine them into new items, which can then serve as components for even newer items.

However, an easily argued counterpoint to this theory is: the more semi-finished LEGO pieces we have, the less likely we are to assemble them into something meaningful, thus increasing the cost to create new, meaningful objects.

This reasoning, which contradicts the technological optimists, aligns more with our everyday intuition—that it’s easier for a poor student to improve from 1 to 60 points than to go from 0 to 1 or from 99 to 100.

If we adopt this more pessimistic prediction about the next technological revolution, it compels us to use traditional means—such as laws, culture, morals, economics, etc.—disdained by technological optimists, to address the issues technology brings.

But first, defining the problem is a prerequisite for solving it.

Before concluding our analysis of internet technology, we must understand in which areas the internet’s attempts have failed—or in other words, which promises the internet has not fulfilled.

The early internet did bring welfare to society as promised, yet some goals remained unachieved or even backfired. This is largely because technological optimists misjudged technology’s leading role in social change and underestimated the impact of societal factors on technology.

Thus, the next chapter will explore the unfulfilled promises of the internet, examining which aspects of the interaction between the internet and society were determined by technology and which were shaped by societal factors.

Only then can we discern the balance between technological and societal influences in their ultimate impact on society.

2.2 Failure to Narrow Social Gaps

On April 5, 2005, the book “The World Is Flat” was published in the United States, and two years later, it filled every bookstore in China and the stands of pirated book vendors on overpasses.

The book depicted a world becoming flat, closely interwoven through networks of trade, finance, and information, with the internet viewed as the final piece that made this flat world a reality, leaving no barriers to prevent humanity from connecting across vast geographical expanses.

The promise that the internet would flatten the world deeply imprinted itself in many people’s minds.

This promise was based on a natural assumption: if you could access dollars through financial networks, purchase goods from the US through trade networks, work for American companies, browse American information and receive American education via the internet, you were, to some extent, American, and vice versa.

However, this assumption was incorrect, entirely overlooking the centuries of cultural and economic disparities between regions.

Like the global financial and trade networks, the internet, as a connector of information, merely meant connecting two ends of the network without necessarily balancing them.

Just as the opening of global trade historically led to the colonization of Asian countries like India and China by European powers, and the opening of financial networks laid the groundwork for global economic crises, the connection established by the internet did not automatically level informational disparities but indeed created more opportunities to exploit these disparities, with some evolving into new forms of exploitation based on information.

Here, we explain the above from three different levels.

The first level is quite straightforward: the “infotainment” issue faced on a micro-level.

The theory of “infotainment” itself won’t be detailed here, as I don’t agree with all its viewpoints. It’s referenced to denote “free, easily accessible, and highly entertaining forms of entertainment.”

The claim that “infotainment” leads to a societal downfall is a fallacy, but “infotainment” indeed contributes to informational disparities.

In the mobile internet era, easily accessible entertainment products fall equally into everyone’s hands, but it’s those with lower digital literacy who are more likely to become immersed. For these individuals, the internet serves to increase informational disparities rather than bridge them.

Free and easily accessible entertainment occupies time they could otherwise use to learn and advance themselves.

In contrast, for those with higher digital literacy, these entertainment products are just a part of their online life. They know how to manage their time and internet behavior better, using the internet to access more information, knowledge, and opportunities for personal advancement.

Simply put, urban youths can use the internet to learn programming, job hunt, create Vlogs, learn foreign languages, and network with influential people. For rural youths, factory workers, and migrant laborers, the internet is often reduced to brainwashing pop songs, risqué live streams, low-quality games, illegal offshore gambling sites, and payday loans.

The issue of “infotainment” is not a universal problem; it’s a subset of the broader issue of informational disparity brought about by the internet. The biggest problem is that previously no one anticipated the internet could increase rather than decrease informational disparities.

In both the West and the East, the idea that “internet proliferation increases informational disparities” was unforeseen, and many still believe today that “the ability to connect to the internet” is the most critical factor influencing informational disparities and the digital divide. To address this, governments, charities, and business organizations worldwide have supported massive internet infrastructure projects and distributed free devices (smartphones and computers).

A 2018 Pew Research study showed that nearly 100% of white, black, and Hispanic American teenagers have access to smartphones1, but this has not positively impacted their academics or employment prospects. Not just smartphones, even PCs, which have “become” tools of productivity, show no direct effect on employment or education outcomes2.

Beyond Pew’s research, the Brigham Young University’s Center for the Study of Elections and Democracy (2019) 3found that 31.9% of teenagers from divorced or single-parent households spend more than three hours daily on screens, compared to 15.1% from two-parent households. Among married white households, 54.7% of teenagers spend less than an hour daily on screens, while in similar black households, the figure is 28.4%. In white two-parent households, 17.2% of teens use social media for more than 1.5 hours daily; in black two-parent households, this rises to 32.1%.

These findings are corroborated by similar studies by Common Sense Media in 20194 and Northwestern University in 20115.

Thus, for people already in economically disadvantaged positions, the internet’s positive effects are far less significant than its negative impacts.

Another more intuitive observation is in China, where no strict anti-addiction system can effectively govern “left-behind” children without direct parental supervision. Essentially, this falls into a confrontational mindset, where no company, platform, or even government can win the “battle” against millions of children—only their parents can.

This gap in usage patterns of the same medium, known as the “third level digital divide,” was first proposed by Jan van Dijk, a prominent sociology professor at the University of Twente, in 2002. This concept focuses on the self-constructs created by different societal groups’ internet habits, even when they can use the internet equally fluently (having crossed the first two digital divides). Many studies have confirmed that the internet plays a more positive role among relatively affluent populations, while its impact is more negative among poorer populations that can access the internet. Further citations are omitted here for brevity.

On the meso level, the internet has established foundational conditions for exploitation based on informational disparities, a phenomenon starkly evident between the United States and Europe.

The borderless nature of internet services and goods means that for the past 30 years, most countries have not had an “internet customs” barrier. This has led to a situation where internet companies from one region can operate in another, bypassing taxes, laws, and economic policies that are the pillars of modern societies.

To illustrate with an imprecise example, traditionally, any company wishing to advertise to European users would have to conduct its advertising transactions within the EU, as advertising entities needed to own media assets (such as newspapers, TV stations, billboards, etc.) within European countries. With the occurrence of advertising transactions, these companies were obliged to pay taxes as per the regulations of each country.

In the internet age, however, any company can target European users with ads through Facebook and Google from any location. The actual transaction could occur where the advertiser is based (via a local Google agency), or default to the United States, where Google or Facebook are headquartered.

This means that a Chinese company advertising in Europe generates GDP and related tax benefits for China or the United States.

This is more sophisticated than any tax evasion scheme in physical trade, as it essentially disconnects a sales activity targeting European citizens from Europe in physical space.

Although taxes are not always a popular topic in public discourse, it’s crucial to reiterate that most countries on Earth still rely on tax revenue to provide security, welfare, and infrastructure to their citizens.

Facebook and Google can easily harvest the attention of European users, creating economic value that primarily benefits regions outside Europe. Whether or not this is the intention of internet companies, it objectively constitutes a new form of economic colonization in the digital age, cutting off a source of funding for social welfare in EU countries, with the citizens of these countries largely unaware of this dynamic.

The essence of this phenomenon is that the operations behind internet services and goods are still largely run by physical companies.

The “modern corporate system” born in the early stages of the capitalist revolution is no longer modern; it is limited to the tangible, atom-based economy, with strong trade network attributes. The economic and tax systems designed for modern enterprises are equally outdated, preventing the economic benefits of the internet from penetrating alongside its products and services.

This situation mirrors the emergence of global trade networks centuries ago, where everything capitalist countries did for the rest of the world post the first and second industrial revolutions was beneficial at the micro level.

From a politically incorrect standpoint, colonialism opened up markets in underdeveloped countries, allowing consumers there to access higher quality and cheaper goods. The spread of trade networks also facilitated the influx of advanced ideas into these countries, eventually triggering the capitalist revolution that overthrew most feudal dynasties worldwide.

However, this cannot negate the immense suffering experienced by countless “backward countries” and their people through the opium trade, slave trade, aggressive wars, and civil strife during this historical process.

Though from a broader perspective of human history, the original accumulation of capital is seen as necessary, and its specific historical stage sometimes justifies it, this doesn’t mean we should aim to repeat this process in the 21st century.

The EU has been actively trying to address this issue in recent years, introducing regulations such as the General Data Protection Regulation (GDPR) and the Digital Single Market Copyright Directive (COD). However, these efforts have not been very effective so far. The GDPR, in particular, has impacted EU’s local internet businesses more severely than multinational giants, effectively becoming a self-inflicted wound. Finding a suitable balance between trade protectionism and economic colonization remains an unanswered question.

On a macro level, the speed at which risks spread through the internet far exceeds that of economic benefits and welfare, which will long-term affect the form of human social governance and may even alter the course of our history. This issue will be discussed in detail in Chapter 4, “The Last Straw in a Risk Society.”

2.3 Failure to Break Down All Barriers

In 1992, American political scientist Francis Fukuyama published “The End of History and the Last Man,” positing that with the dissolution of the Soviet Union and China’s initiation of reform and opening up, the essential conflicts between human societies had vanished. According to Fukuyama, human civilization, after a long and painful evolution, had finally reached a terminal mode (which he identified as the American model). This meant that, having found this optimal mode, human society would cease to evolve, and conflicts and wars between nations, regions, and races over social paradigms would entirely disappear.

For nearly 30 years following the publication of “The End of History,” Fukuyama’s thesis occupied a significant place in Western mainstream discourse, while being critiqued by countries represented by China in East Asia, thus frequently revisited. However, its predictive and summarizing power as an academic work did not match that of its academic rival, Samuel P. Huntington’s “The Clash of Civilizations and the Remaking of World Order,” published the following year.

Huntington’s “The Clash of Civilizations” contested the disappearance of conflict and the end of history, predicting that as political struggles conclude, conflicts on the levels of civilization and culture would commence. These conflicts, more complex and latent than those of the past, would ultimately manifest in familiar forms of economic, financial, and physical warfare.

Civilization, though difficult to precisely define, objectively influences every individual’s self-identity and behavior. Huntington noted that people define themselves through ancestors, religion, language, history, values, customs, and institutions, preventing us from making Pareto optimal decisions in every situation as the rational economic actors assumed in economics.

This deep-seated influence, from micro to macro levels, leads to nations making starkly different top-level decisions even under the same political and economic systems and shared interests.

The clash of civilizations is self-reflexive, meaning its competition does not produce an “optimal solution” like the competition of social forms might. This necessitates clarifying the difference between multiculturalism and cultural pluralism.

Taking the United States as an example, it is recognized as the world’s largest melting pot, its inclusivity of multiculturalism surpassing other nations. This allows members, goods, or cultural symbols from any civilization or culture to survive and grow to some extent within American society.

However, the result is a single culture—a multicultural American culture—where other cultures are mere segments of a polyphonic narrative under the dominant American cultural narrative.

In other words, multiculturalism refers to a culture that can accommodate many elements from different cultures; essentially, it is one culture. Cultural pluralism, in contrast, refers to multiple cultures; it does not require each culture to be inclusive of others, ultimately converging into one culture.

The reflexivity mentioned earlier implies that advocating for cultural pluralism requires tolerating some extremely incompatible cultures; whereas promoting multiculturalism essentially involves diluting the integrity and independence of other cultures.

So, what impact does this reflexivity have on the internet?

Intuitively, the internet seems to support Fukuyama’s “End of History” thesis, suggesting that as political and economic systems increasingly converge, the internet facilitates communication that could lead to the merging of global cultures into a unified multicultural narrative.

Contrarily, due to the inherently inefficient and disorderly nature of individual communication, the internet has exacerbated stereotypes and discrimination between different civilizations. Cultural symbols are mistaken for the culture itself in interactions between different groups, leading discussions to become more black-and-white.

To depoliticize the explanation: why are group chats and BBS forums often more engaging, kinder, and produce higher value content compared to platforms like Twitter or Weibo?6

The reason lies in the specific themes and discursive systems of group chats and BBS forums, which allow for smoother and deeper discussions on certain topics. Low-quality or irrelevant contributions are quickly dismissed within the community.

Conversely, on platforms like Weibo, the crossing of discursive systems means there’s little common ground for meaningful dialogue. Especially in debates, offensive language often proves more effective than factual arguments—this is colloquially known in the Chinese internet sphere as being “dragged down to the level of an idiot and beaten with experience.”

This represents a reverse selection process, where in a relatively narrow public sphere, more rational voices can dominate; whereas in an absolutely open environment, whoever is louder is “more rational.”

This eventually leads to any culture being represented by those least capable of showcasing its virtues, forming the impression others have of that culture.

Additionally, the presentation of communication on the internet poses certain issues. Fragmented, visual, and random information dissemination does not aid in building a rational understanding of one another.

Whether we align with Fukuyama’s end of history or Huntington’s clash of civilizations, what we undoubtedly do not wish to see is this level of quarreling. The competition between civilizations, whether it results in convergence or diversity, should be conducted more gracefully rather than through mutual disparagement.

At this moment, we might almost be grateful that AI translation has not advanced to the point where every person on Earth can converse fluently regardless of language barriers. Otherwise, logging online would expose any individual to the opinions of 7 billion critics.

This phenomenon will profoundly impact all aspects of our society, especially after internet products have effectively securitized and commodified public attention. Rational discussions hold significantly less commercial value compared to traffic-generating conflicts.

The commercial nature of the internet exacerbates this trend. While the internet may not efficiently transmit culture itself, it serves as an excellent conduit for the commodification of cultural symbols. Anyone can purchase kimonos, African-American music, or Chinese knots online, leading to these commodified cultural symbols increasingly representing the cultures that produced them.

However, cultural symbols are concrete, limited, and static; they are easily misused, misunderstood, and attacked.

The phenomenon of using symbols to represent groups can even corner the concept of free speech.

The 2019 incident involving Houston Rockets’ General Manager Daryl Morey’s tweet, which was considered insulting by China, is a focal example of this issue.

Before this incident, the NBA had experienced at least four other controversies deemed offensive by China, leading to two official broadcast suspensions. Yet none had sparked as widespread a reaction in both the US and China as Morey’s comment.

On one hand, this was due to an increased sense of collective honor in China. On the other hand, most Chinese people were unaware of the first four incidents.

The act of insult requires an object to receive the action. Before the internet, insulting a country was a relatively vague act. A “private” insult, if private enough, might not constitute an insult at all.

This is not an ostrich policy; even in daily life, disdain or mockery within one circle towards another exists. It’s akin to “someone bad-mouthing someone else behind their backs in a small group”; although unethical, it doesn’t have a significant impact, and the “insult” doesn’t materialize because the insulted party is unaware of it.

However, with the internet, public social media platforms have for the first time given individuals the ability to insult every member of a group.

This marks the significance of the Morey incident; his comment made nearly every Chinese person who saw it feel insulted, making subsequent reconciliation extremely difficult.

The abstraction of group symbols on the internet into individual attacks provides targets, while the speed and method of internet information dissemination offer the tools. In depoliticized daily discussions, mutual attacks and denigration between fan groups are also realized through this process.

Going back to basics, does freedom of speech imply the right to express unprofessional, irrational, or even inaccurate content? In the era when the term “freedom of speech” was coined, it did because in a complex modern society, expecting every individual to speak only with full information, professional knowledge, and absolute rationality would effectively silence everyone.

But in the era when the concept of freedom of speech emerged, individuals were not endowed with the simple, barrier-free capability to reach out to 7 billion people.

Or to put it another way, is “ensuring that no speech is criminalized” mutually exclusive with “the ability to reach out to 7 billion people”?

2.4 Failure to present the real world

Twitter, launched in 2006, is undoubtedly a milestone in the history of media. Unlike Facebook, Twitter pioneered the “public square” style of social networking, which has revolutionized the media ecosystem of every country and region around the globe over the following decade.

Both Twitter and its counterparts like Weibo have, for the first time, given individuals the right to speak directly to the public without any intermediary institutions. This birthed the concept of “independent media” or self-media, which at one point posed a threat to traditional media worldwide.

Embedded in Twitter’s posting prompt, “What’s happening?” is a strong metaphor. Initially, most tweets indeed contained the six elements of news reporting: What, Where, When, Who, Why, and How.

This led to a simplistic illusion: no journalist could describe an event faster, more accurately, and more truthfully than those directly involved in it.

This evolved into a promise from new and self-media to the public, using it as a weapon to compete with traditional media across the globe. For a while (until around 2015), in the eyes of the public, self-media was seen as a more authoritative source of information than “institutional media.”

The naive fantasy stems from the idea that the assault of public square social media on traditional media would come from controlling firsthand sources.

To get the most accurate reports from the scene as quickly as possible, traditional media spent hundreds of years establishing thousands of journalist stations worldwide, incurring significant human resource costs. For Twitter, every user acts as a correspondent without any cost to the platform.

Despite the ongoing trust crisis in traditional media, both in China and the West, it must be acknowledged that even those outlets often criticized as “fake news” possess much better fact-checking mechanisms than social media.

For instance, some discredited online self-media are mockingly referred to as the internet versions of “Story Meeting” or “Listener,” magazines known for their human interest stories. However, “Listener” magazine has always employed fact-checkers independent of its editorial department to verify key details in interview materials. Meanwhile, “Story Meeting” has always been a publication focused on original fictional literature.

The more critical issue is that for every user of public square social media, there’s no obligation to ensure the accuracy or truthfulness of their statements, an impossible expectation for any individual.

Especially for significant events, the self-reporting encouraged by public square social media can lead to even more irrational states. This is easily supported by psychology, such as witnesses of a shooting mistakenly identifying an unrelated person as the shooter or fabricating what was said at the scene.

Victims of natural disasters, such as earthquakes, might project their frustrations and the losses they’ve suffered onto human-made disasters due to years of irrational dissatisfaction.

Considering social media as a whole, and taking into account that only about 10% of Chinese netizens have a bachelor’s degree or higher7, along with a potential mental disorder prevalence rate exceeding 17% in the Chinese population8, global social media might be one of the least reliable media forms in human history.

Moreover, as almost every incident involves different individuals, personal users are not held accountable for their long-term news credibility, creating opportunities for those intentionally spreading falsehoods. Currently, it’s challenging to discern whether the virality of news subjects stems from the individuals themselves or from agenda-setting teams.

In my analysis “Why do we feel more and more media lack conscience?”, I explained that from a micro perspective, traditional media might not necessarily report the truth, as the sense of reality often holds more importance than actual truthfulness.

However, compared to traditional media, personal media on public square-style social networks often offer ten times the sense of reality with only a fraction of the truthfulness, which is undoubtedly harmful to society.

Fundamentally, public square-style social networks endow everyone with media attributes, damaging both media integrity and social interaction. This forces us to tread carefully in our normal social behaviors, which should not have to bear public responsibility, yet it significantly weakens the credibility of serious news dissemination.

In other words, serious media can’t compete with individual speakers in terms of “sense of reality” and “speed,” while individuals, due to their media attributes, are pressured to ensure their statements possess “truthfulness” and “professionalism.”

Public square-style social media is not the only product form causing damage to the media industry; recommendation algorithms play a similar role.

In traditional media or the web media era 1.0, editors not only reviewed submissions but also played a crucial role in placing the right content in the right spot to capture readers’ attention. This authority to edit submissions was foundational to their ability to discard low-quality pieces, remove redundant information due to space limitations, and rank news stories and events.

However, in the era of personalized algorithms, editors no longer control the layout. This means that both traditional and new media editors are relegated to merely exercising their baseline authority to review submissions, mechanically adhering to editorial standards to ensure the safety of published content. This has led to a long-term decline in editors’ status within the media industry, which is partly why the quality of media content has been steadily deteriorating.

The sentiment “Who are you compared to what the masses like?” challenges the mainstream view that editors should have the authority to set agendas. However, as discussed earlier regarding the digital divide, when digital literacy falls below a certain threshold, people tend to select lower-quality content to satisfy short-term dopamine spikes, something non-human AI editors cannot counter.

From another perspective, personalized algorithms unify diametrically opposed viewpoints under a nebulous consensus, exacerbating biases and stereotypes over time.

Taking the media ecosystem in the United States as an example, traditional media used to display a clear left-right divide, with left-leaning media like The Wall Street Journal, CBS, and The New York Times and right-leaning outlets like The Washington Post and FOX offering different interpretations of the same events. This dynamic maintained a state of chronic confrontation between the two stances.

In the new media environment, it’s challenging to pin down the political stance of Google News, Apple News, or YouTube. They present stories that align with the reader’s preferences, essentially showing readers what they want to see and suggesting that this represents the entire world.

During the 2019 Hong Kong protests, some parents in Hong Kong noticed that the content displayed by YouTube varied significantly between their accounts and their children’s accounts. Assuming Google had no intervention in this process, this outcome aligns with what a “neutral” recommendation algorithm aims for—showing readers what they want to see, suggesting it’s the entirety of global perspectives.

Previously, even someone who subscribes to The New York Times might, during significant events, actively seek out FOX News to hear the “other side.”

However, under algorithm-driven media landscapes, while opinions from both ends of the spectrum still exist, they no longer have a battleground for confrontation. Different news stories are distributed to entirely different audiences, creating isolated and increasingly extremist opinion chambers.

This separation occurs across internet products driven by algorithms, not just within media apps. On platforms like TikTok, Kuaishou, and Weishi, we see many groups that never intersect, and each group finds a vast audience. For instance, when discussing live-streamed sales, many think of Li Jiaqi. However, unlike Li Jiaqi, who targets first-tier city markets, Kuaishou influencer Xin Youzhi sells brands and products unknown to urbanites, achieving far greater success with his down-to-earth style.

This phenomenon of people dividing into groups isn’t new to the internet era, but the internet, especially recommendation algorithms, has indeed accelerated it.

In the pre-internet era, media agenda-setting was an open secret. It was commonly understood that each media outlet would edit material with a bias according to its stance. This, however, meant that one could easily seek out media with an opposing agenda to understand “the other side” of the story.

The internet has stripped both new and old media of their agenda-setting power, but it hasn’t dissolved agenda-setting itself. Instead, it has solidified agenda-setting within each individual’s world through information echo chambers.

Unlike traditional demographics, algorithm-driven segmentation makes it easy to overlook the world beyond one’s “group.” The cocoons of information and social interaction make those inside more prone to misconstruing their view as representative of the entire world.

Search engines were the last bastion for breaking out of these information cocoons. However, with the global decline in search engine usage, who knows how long it will be before recommendation algorithms take over search engine rankings too?

When people’s method of accessing information shifts from actively subscribing and searching to passively “pulling to refresh,” the internet ceases to be a window to the world and becomes merely a flattering mirror.

Put bluntly, if someone consistently reads articles with 100,000+ views, they might assume these represent the mainstream opinion. But in reality, the algorithm is just pushing articles that align with its “taste.” For the over 800 million internet users in China, even a public account consistently hitting 100,000+ views is merely tapping into a tiny niche within the entire discourse, not representing the mainstream viewpoint.

For every speaker and audience member within these discourse arenas, there’s an increasing belief that “what I write/read reflects the majority’s view,” leading us further away from understanding the real world.

If we can’t even recognize the true nature of one another, the hope of achieving mutual understanding becomes even more elusive.

3. The Fall of the “Meta-Narrative”

3.1 The Most Important Meta-Narrative of the Past Century: Globalization

Metanarratives are overarching narrative structures or discourse systems that attempt to provide definitive answers to humanity’s past, present, and future social activities.

They are characterized by their ability to address societal questions far beyond the era of their inception, offering seemingly plausible explanations. Examples include Christianity, democracy, and communism, each a metanarrative born in different historical periods of human history9. These examples align with the definition highlighted.

But are metanarratives inherently bad, false, or unnecessary? Not necessarily.

Yuval Noah Harari, in his bestselling book “Sapiens: A Brief History of Humankind,” attributes humanity’s greatest ability to “create and believe in fiction,” a capability even more crucial than our tool use.

A credible and positive metanarrative can become a self-fulfilling prophecy to some extent, even if the methods behind it are fabricated and its original purposes greatly diverge from what is outwardly stated.

Taking globalization as a direct example can help understand the role of metanarratives more clearly. The behavior of globalization originated in the late 19th and early 20th centuries, but the concept (narrative) of “globalization” was only coined in academic circles in the 1960s10.

The United States undoubtedly led the initial push for the “globalization” metanarrative, bolstered by wealth accumulated during World Wars I and II and subsequent military power. The narrative’s original goal, aside from uniting non-communist countries, was also to construct a form of “post-colonial global hegemony.”

This narrative eventually achieved the initial effect its narrators intended: post-World War II America managed to control the global economy without producing goods themselves, using a new form of economic leverage with the dollar. This allowed for cross-regional exploitation and economic colonization without bloodshed, placing the US in an unchallengeable position for a long time.

However, it’s important to note that globalization not only fulfilled its narrators’ hidden agendas but also genuinely delivered on its promised benefits: increased interconnectivity, mutual understanding, mutual benefit, and reduced warfare.

However, globalization also delivered on its explicit promise of mutual exchange, understanding, mutual benefit, and reduced warfare. While the distribution of benefits during globalization has been uneven, increasing disparities between economically disadvantaged and advantaged areas, even regions and nations in the “exploited” role have seen significant development compared to their previous states.

Nuclear deterrence may underpin globalization, but without the “story” of globalization, nuclear weapons alone couldn’t have brought about the welfare for all that globalization has—truth may only extend as far as cannon range, but guns cannot replace Alipay in facilitating every transaction worldwide.

Meta-narratives serve a vital purpose by crafting fictional yet appealing stories that motivate us across time and space, facilitating cooperation between groups that are strangers or even have conflicting interests. While we may never reach the “happily ever after” depicted in fairy tales, it doesn’t prevent everyone from benefiting throughout the process.

The internet is undeniably part of the globalization meta-narrative, originating from the Cold War era’s impulse to unite the world (the Soviet Union also had a computer network plan for the communist bloc).

While the internet has realized globalization, it eventually became globalization’s own antithesis.

As previously discussed, the internet has produced effects in key areas that are contrary to its initial promises, revealing the reality behind the “globalization” story.

After completing the logistics and financial networks with the information network, we’ve achieved the interconnectedness that globalization described. However, we haven’t attained the perfect world promised in the globalization narrative. With everything now connected, the story of “connecting everything” no longer inspires.

From a more pragmatic standpoint, if the conclusion about the end of the internet technology revolution described in section 1.2 is correct, it will directly affect the macroeconomic cycle, and economic downturns historically bring dark times for humanity.

The current era’s darkness is characterized by the rise of nationalism, populism, conservatism, and digital authoritarianism.

Internet technology, being a part of globalization, faces obstacles due to the disintegration of globalization itself (as mentioned earlier, attempts by various countries to sever connections, impose taxes, and restrict content spread). The dilemma the internet faces is akin to a serpent biting its own tail, unable to solve its problems independently.

However, the internet’s disruption extends beyond just the “globalization” meta-narrative; its long-term, irreparable damage to the very tool of social construction that meta-narratives represent is of greater concern, as discussed in section 2.3.

It’s crucial to note that while meta-narratives can provide answers to all human social activities, they are not meant to guide all such activities. If this concept is difficult to grasp, consider the periods in human history when the dogmatic enforcement of Christianity, democracy, and communism led to significant consequences.

This text itself is a meta-narrative, aiming to offer a conclusive analysis of the internet industry, explaining phenomena that traditional tech media and business commentary fail to address at a micro level.

Thus, this paper(you are reading) doesn’t offer specific guidance on a micro level but describes a possible trend in the internet industry and its impact on current society.

Every internet company, regardless of size, should prioritize the micro-level issues specific to its operations. Decision-making tools at the micro-level hold more practical significance than the narrative framework provided in this text.

If you’re reading this seeking growth strategies for the next quarter, unfortunately, this text is unlikely to offer direct assistance. Apologies for the time spent, but thankfully, you’re halfway through.

3.2 When “love” is no longer love

In 2016, MIT developed a research tool named “Moral Machine,” which aimed to conduct a social experiment by presenting global internet users with a series of scenarios. The project modernized the traditional “trolley problem” into an online survey, where participants were asked to assume the role of an autonomous vehicle forced to choose between two options in a split-second decision scenario.

The survey included various controlled variables, asking participants to make choices between hitting a man or a woman, a human or a pet, a white person or a person of color, an older adult or a younger person, a wealthy individual or a poor one, among others. Nearly 40 million choices were collected from respondents in 233 countries.

The comprehensive results of this study were published in “Nature” in October 2018, revealing some findings that align with our general understanding, such as the tendency of people to save more pedestrians when given a choice.

However, when the choice did not affect the number of casualties, participants showed clear biases that contradict the moral standards we hope to impose on AI. For instance, there was a preference for sparing the lives of wealthy individuals over poor ones, thin people over overweight ones, women over men, and younger people over older ones. Interestingly, within the older demographic, there was a tendency to spare older men over older women.

This demonstrates a form of hypocrisy in public morality, undermining the legitimacy of demanding moral reasoning from algorithms—if humans themselves struggle to define “correct” moral decisions in emergencies, how can we expect machines to follow a fixed moral standard?

The experiment highlights just a small aspect of how algorithms challenge human morality. The broader internet culture’s pursuit of precision is unraveling more narrative elements, making the crisis of metanarratives we discussed in the previous section distinct from past shifts in dominant narratives.

A major metanarrative is typically composed of many fragmented narrative elements. When one collapses, a new metanarrative emerges, often constructed from selected elements of the old narrative combined with new logic.

However, the crisis that the internet poses to metanarratives is not about ending a specific narrative. The blurring boundaries brought about by digital technology lead to the disintegration of all fragmented narrative elements, fundamentally destroying our trust in any form of narrative.

Taking love as an example, it’s arguably the most pivotal narrative element, intricately woven into the fabric of human history’s every metanarrative.

From the myth of Pangu splitting heaven and earth to the Biblical command “Let there be light,” narratives of love have been central, spanning the divine love of gods for humanity, the stern love between monarchs and subjects, the respectful love of citizens for their rulers, the familial love of parents for their children, and the romantic love between individuals. Even the Declaration of Human Rights begins with a reclaiming of universal love from the hands of the ruling classes.

Love’s power lies in its universal appeal, its ability to stir the deepest emotions inherent to our species, creating a broad consensus.

Consider an extreme scenario: a victorious soldier hesitates to fire upon wounded or non-combatant enemies, moved not by abstract humanitarian principles but by a visceral response to the immediate context—groans of agony, expressions of pain, the smell of blood and gunpowder, the desolation of the battlefield under a bleak sky.

Abstract, verbalized, digitized notions of humanity often fail to drive such instinctual decisions.

Conversely, if soldiers operated drones or robots from thousands of miles away, pressing a “clear the remnants” button in a command center, the compassionate love that might restrain their hands in person plays no role.

Abstracted and verbalized love is not the love that commands widespread consensus. Regrettably, the internet is fundamentally a network of information, and any discussion of love on it represents only an abstract form of love.

As noted previously, individuals from distinct civilizations lacked extensive platforms for dialogue, allowing people with different cultural backgrounds to tacitly accept each other’s definitions of love without conflict. Typically, it was easier for people then to reach a consensus that someone is capable of love, despite any conflicts of interest or belief.

Now, we see increasing disagreements over the definition of love online, leading to arguments, attacks, and even extremism. On the internet, love has fallen from its pedestal, becoming a casualty in every secular conflict, losing its unifying force. Notably, conflicts like those between religious love and love in same-sex relationships are prime examples.

This deconstruction is partly due to the excessive and simplistic connections described earlier and partly due to the internet’s precision in definition and memory. In simpler terms, the internet doesn’t just “dig up dirt” on public figures and “stigmatize” them; it does this to every narrative element.

According to a study by The Atlantic11, in 2019, 35% of Republicans and 45% of Democrats opposed their children marrying someone with different political, religious, or racial views, a significant increase from just 5% in 1960.

While the internet isn’t the root cause of these divisions, it acts as both an accelerator and a cementing agent for them. Even today, we can still find the most venomous tweets from the mutual vilification among U.S. netizens surrounding Trump during the 2016 election being circulated and used for further attacks.

We remember how we initially supported charitable organizations, only to find out they were a sham. We recall our initial support for environmental causes, which later turned into profit-making ventures by some organizations. We also remember how animal protection efforts went from being noble to radical.

Individual humans forget, and so do collectives, but thus far, the internet almost never forgets. Years later, when you decide to engage with charity work again, search engines and social media tirelessly remind you how the incident with Guo Meimei a decade ago led you to decide “never to engage in charity again.”

You revisit the decision-making process and are astonished to find the internet has not only precisely preserved the outcome of your decision but also all the materials that informed your decision, “ensuring” you won’t give the other party a second chance.

Every old narrative element remains tarnished by historical blemishes, unable to redeem itself. Every new narrative element, in its rise to public recognition, triggers massive, unconstructive discussions.

Ultimately, we might be heading towards an era devoid of narratives.

3.3 Why Have States, Corporations, and Organizations Lost Trust?

Entering an era devoid of any grand narrative might seem like a step towards pure rationality, but is that necessarily a good thing?

Firstly, this is a profoundly long-term question. Since the dawn of civilization, humanity has continuously believed in various grand narratives, and any conclusions drawn in the short term are unreliable.

However, the absence of grand narratives is likely to have a negative short-term impact on our current social formations, as a significant part of human society has historically relied on these narratives to function.

As early as 1970, French scholar Jean-François Lyotard in “The Postmodern Condition: A Report on Knowledge” identified “grand narratives” as the hallmark of modernity. The dissolution of grand narratives signaled the onset of postmodernity, a period that corresponded with the “Beat Generation” and “hippie culture” in the United States, with China only expected to encounter this “affluent problem” among generations born after 2000 or 2010.

The objective benefit of grand narratives to human society is not to provide us with a manual for guiding every detail of our lives but to foster belief in the future, in each other, and in our commonalities to facilitate large-scale social cooperation.

To simplify, even for most modern citizens with higher education, understanding how systems like modern transportation, finance, or healthcare operate is often beyond their grasp. Approaching these systems with a skeptical, inquisitive mindset could significantly reduce the conveniences brought by modernity.

For example, a patient’s willingness to allow a stranger, a doctor, to inject medicine into their veins is not based on a comprehensive understanding of the drug’s molecular structure or pharmacokinetics. Instead, it’s based on trust in the scientific safety and efficacy evaluation mechanisms that the drug underwent before market release, and that the doctor prescribing it has undergone systematic professional education and assessment.

Often, our ignorance is second-order: most patients do not fully understand what “scientific drug evaluation” and “scientific medical training” entail, but they place faith in the scientific nature of their treatment (even though these treatments are indeed based on science).

This form of “faith” is beneficial. The complexity of modern society means its operations are far beyond the comprehension of individuals and even small or medium-sized organizations. Once we engage with modern society, we inevitably interact with areas beyond our knowledge (even if you’re not a doctor, you still require medical treatment). Therefore, trust in people, institutions, or systems that specialize in these areas outside our expertise is essential for the functioning of modern society.

This leads to a rather alarming phenomenon: the modern society driven by technological productivity operates on the basis of a widespread “superstition” among the majority of its members. Evidently, the narrow definition of trustworthiness (including governmental credibility, corporate trust, social organization integrity, and interpersonal trust) struggles to keep pace with the rhythm of the internet era. Meanwhile, the internet endows us with an unparalleled capacity for skepticism. A modern conspiracy theorist’s article, fuelled by extensive internet research, might appear more “credible” than a serious scientific paper from a century ago.

A prime example of this model is the story told in “Bad Blood: Secrets and Lies in a Silicon Valley Startup,” a piece of investigative journalism that won the Pulitzer Prize in 2019. It chronicles the tale of an entrepreneur, almost entirely lacking in medical background, who carried out a massive medical device scam over 12 years in what is considered the country with the strictest healthcare regulations in the world.

From the company’s foundation in 2003 following the entrepreneur’s departure from Stanford, to the first systematic exposure by The Wall Street Journal in 2015, the scam persisted for over a decade, involving hundreds of thousands of people, without facing significant obstacles from any of the at least five U.S. regulatory agencies that should have intervened.

The precise memory and broad dissemination capabilities of the internet mean that the impact and speed of credit damage are much greater than in previous eras. This issue is particularly pronounced in the realm of government credibility.

To illustrate with an abstract political example:

  • Assume Country A has a population of 30 million, and Country B has a population of 1.5 billion, or 50 times larger.
  • Assume Country B’s land area is also 50 times larger than Country A’s.
  • Assume both countries have equal crime rates.

In the pre-internet era, we would conclude:

  • Given equal crime rates, the net number of crimes in Country B should be 50 times that of Country A.
  • Since Country B’s land area is 50 times larger and the speed of news dissemination is geographically bound, the likelihood of an individual in Country B encountering a murder report in the local media daily is equal to that in Country A.
  • Citizens of Country A and B would perceive their countries to have similar levels of public safety.

In the internet era, it changes to:

  • Given equal crime rates, the net number of crimes in Country B should be 50 times that of Country A.
  • Since online media often operates within linguistic boundaries and can push a significantly higher volume of news daily, any murder case in Country B can be broadcast nationwide.
  • An individual in Country B will see 50 times the crime reports compared to someone in Country A.
  • People in Country B feel their country is incredibly unsafe.

Previously, public safety incidents were evenly distributed across Country B’s land, which is 50 times larger, and disseminated through local newspapers, radio, and TV, impacting only those citizens related to or geographically close to the incident. Only when a public safety incident escalated beyond local resolution did it attract attention from citizens in other areas.

In the context of the internet, where the speed of information dissemination has increased and barriers have been lowered, any public safety incident in Country B becomes national news, capturing the attention of all its citizens. Due to people’s strong empathetic response to safety issues, the public tends to perceive incidents occurring in other regions as directly relevant to themselves. This could lead to disregarding the fact that Country B’s crime rate might have always been low and comparable to that of Country A, resulting in a perception of deteriorating public safety in Country B at a rate 50 times faster than in Country A.

If you think this example doesn’t reflect reality, unfortunately, there’s a real-world instance:

2018 was a disastrous year for Didi, the ride-hailing company, which faced several severe safety incidents involving its drivers, despite carrying 20 billion passengers annually.

Mentioning Didi’s passenger numbers isn’t to praise the company but to illustrate the scale comparable to the “50 times population” concept mentioned earlier. Didi’s national carriage volume in 2018 was dozens of times higher than the total of all traditional taxi companies combined.

Looking at the cold numbers, during the period Didi was embroiled in controversy, the Supreme People’s Court had calculated that the incidence rate of crimes committed by ride-hailing drivers was 0.048 per 10,000 people12, whereas for traditional taxi drivers, it was 0.627 per 10,000. Scientifically speaking, ride-hailing services were significantly safer than traditional taxis.

It’s important to understand that any potential risk, given unlimited attempts, is bound to manifest eventually.

Compared to traditional taxis, Didi’s service is closer to having “unlimited attempts.” Not only was the total volume of rides in traditional taxis significantly lower after 2015, but “traditional taxi” is not a unified brand that the public can easily focus on.

Put simply, each city has several independent taxi companies. A severe incident with a taxi company in one city won’t trigger empathy among people in another city. However, a single incident with Didi in any city catches the attention of potential Didi users nationwide.

It’s foreseeable that if we can’t make the public realize that absolute safety is impossible, Didi will never be released from such accusations. Often, these accusations solidify into a force (usually in the form of excessive regulation), becoming a shackle that hinders further technological development and deprives the technology of its self-healing capabilities.

Thus, we stop believing the “ride-hailing” narrative is a “good story.”

This brings us back to the beginning of this section—the collapse of grand narratives marks the start of modernity’s self-reflection. Although we’ve said that the purpose of grand narratives isn’t to dictate our every action, they are crucial for building consensus and driving us towards a better future.

Belief makes vision possible, while disbelief hinders realization.

3.4 Wrong Answer

In our previous discussion, we delved into how the internet deconstructs grand narratives and hinted at the potential crises that might ensue. Alongside this trend of deconstruction, there emerges another tendency: the codification of morality.

In our unease with increasingly unconvincing grand narratives, a compromise struck between tech optimists and conservatives has been to use technology itself to solidify and emphasize some of our existing narratives, like morality. However, this hasn’t solved any problems; rather, it has potentially placed us in a more passive position.

A clearly defined and powerful set of moral standards (or other grand narratives) is alarming because it often hinders progress in human society on various fronts.

To put it plainly, 2019 was a tumultuous year for artificial intelligence. The public perception of AI plummeted compared to the previous year. We started to grapple with the issues of algorithmic echo chambers, pervasive surveillance cameras, the misuse of face-swapping apps, the exploitation of our privacy by targeted ads, the control exerted by smart classrooms over young students, and the potential for brainwave monitoring to make psychological crimes a reality.

If we attempt to regulate human morality using AI, you can imagine the potential consequences.

In recent years, we’ve enjoyed the conveniences brought by the third wave of AI, despite some modern Luddites opting out upon encountering the invasive “terms and conditions” frequently imposed by tech companies. However, this remains a choice for the few.

In response to public opinion and regulatory scrutiny, most tech companies are on a quest for an AI approach that maintains benefits while minimizing harms. Finding a middle path between privacy and efficiency would be ideal, but if such a path technically does not exist, can we undo the efficiency gains AI has brought to our society in the past five years? The answer is almost certainly no.

This leads to an odd situation: we gradually become accustomed to the “negative” impacts of technology, especially those that conflict with morality. Our children might not see any issue with information bubbles, may view surveillance cameras as symbols of safety, and might believe targeted advertising is perfectly normal, even desirable.

Marx termed this phenomenon “alienation,” while Confucius described it as the breakdown of ritual propriety and music. The “rites” Confucius referred to were the elaborate and hierarchical social norms of the Western Zhou period, which, from our modern perspective, are considered largely outdated. It was precisely because of the breakdown of these rites 2,700 years ago that Chinese civilization had the opportunity to evolve to its current state. As mentioned in discussions of alienation, this phenomenon emerged with the birth of class society and will not disappear until the abolition of private property, class, and state—implying that, for a significant stretch of history, we have no choice but to adapt to alienation.

Thanks to the ambiguity of morality, virtues can be celebrated for millennia without considering that the actual standards behind them might be entirely different. Although each dynasty in China was seen as a period of moral and ceremonial decline by the standards of its predecessors, China has never lost its reputation as a “land of ceremony and propriety.”

Some vague, profound, and national moral values are deeply engraved in our cultural heritage, rather than in the elaborate and rigid rules written down. For instance, the concept of filial piety in the Yuan dynasty differs significantly from today’s understanding.

Even during the Yuan dynasty, not everyone precisely followed the “Twenty-four Filial Exemplars,” but if these stories were coded into AI, they would be executed precisely.

Moreover, the number of humans that AI can influence and control today may exceed the total number of people who have read the “Twenty-four Filial Exemplars” since the Earth’s inception.

This means that an AI, precisely programmed with certain values or morals, would become a self-fulfilling “moral machine.” Like a large nail, it would eternally fix our society’s moral evolution tree in its current state.

While this might be possible, it’s also terrifying, more so than returning the conveniences technology has afforded us.

As surveillance cameras proliferate and personal credit systems are misused, we seem to be moving toward an era of solidified morality. However, as discussed earlier, human morality itself is a narrative, driving us to spontaneously improve in a positive direction.

It is precisely because we can never achieve moral perfection that morality serves as a force for continual self-improvement.

Externalized, enforced, precise morality risks stripping away the capacity for spontaneous goodness, undermining the societal force of morality as a complement to law, and subconsciously pushing more people into the camp of those seeking to exploit moral loopholes.

As Tim Cook has said, “I’m not worried about artificial intelligence giving computers the ability to think like humans. I’m more concerned about people thinking like computers.” Enforcing a current moral code with technological means is essentially making humans think like machines.

“The Essence of Technology” suggests that “the era creates technology, and technology, in turn, creates the era.” When you enter a technology museum, you inevitably see the transformation (or evolution) of human society, culture, and morality brought about by technology, such as chemical flasks, slide rules, ID cards from the Manhattan Project, mannequins in gendarme uniforms, gasoline rationing cards, old cars, and jeeps, etc. All these elements that shaped past societal formations are linked to the technology that was the typical means to an end at the time.

In academic circles, sociologists tend to oppose alienation more than anthropologists do, not out of fear, but because sociologists focus on collective human behavior, and the alienation of individual behaviors is seen as a consequence of societal change, leaving individuals no choice. Thus, alienation is considered an undesirable outcome. Anthropologists, focusing on individual humans, see alienation as a driving force, with societal transformations being the result.

The pursuit of precisely executed morality reflects a certain delusion brought about by the “code is law” concept. Theoretically appealing, “code is law” paints a picture of a world where justice is neither absent nor delayed—an enticing vision indeed.

Yet, even setting aside moral considerations and looking purely from a legal perspective, “code is law” may not necessarily represent the right direction. Legislation is fundamentally a large-scale social contract, resulting in a situation where even with the aid of the internet, establishing or amending laws can take a cycle of 5-10 years across nations.

Intuitively, it might seem that laws should leave no room for ambiguity. However, in practice, the ambiguities of law exist between legislative revisions or before the inception of new laws. For instance, if the legal frameworks governing transportation had been enforced with machine-like precision starting 30 years ago, ride-sharing platforms like Uber and Didi might never have emerged. Instead, we might have seen decades-long transformations of traditional taxi companies into internet businesses, similar to the situation in Japan.

The notion that technology needs to have values really means that the people who write, deploy, and use technology should possess values, not that the technology itself should embody these values. This distinction is crucial in understanding how we can harness technology responsibly, ensuring it serves to enhance human values and freedoms rather than constrain or redefine them in rigid, possibly outdated, terms.

4. The final step into a risk society

4.1 Risks are produced on a large scale in globalization

Entering the last phase into a risk society marks a profound shift in how we perceive and manage societal challenges. The concept of the risk society, introduced by German sociologist Ulrich Beck in his 1986 publication, argues that the risk society will supplant class society as the dominant social structure. In this new paradigm, the main social conflict shifts from class struggle to the struggle against shared global risks, making risk the common enemy of humanity worldwide. Competition among nations increasingly revolves around the transfer and defense against risks.

Beck’s theory posits that risks are indifferent to national borders and social classes. They can spread through the global environment, trade, financial, and information networks, eventually erupting into global crises in the least defended areas.

For instance, environmental pollution and nuclear safety are quintessential examples of mobile risks. Emissions from factories in Europe can cause acid rain in the United States or poison children in Central Asia, with European companies neither able nor willing to address the environmental crises they trigger in non-local regions. Similarly, those affected in other regions have little to no involvement in European industrial planning, leaving them without means for prevention or remediation.

While environmental pollution might represent a long-term and less visible issue, nuclear crises are more immediate and lethal. Despite a seemingly effective nuclear deterrence system developed to prevent the use of nuclear weapons, the peaceful use of nuclear energy presents even greater dangers.

Countries with nuclear capabilities vary in economic development, technological prowess, and political-economic systems, leading to disparate safety standards and measures across nuclear facilities. The global risk of a nuclear leak depends not on the most secure facility but on the least secure one. A single accident at any nuclear power plant could potentially trigger a global extinction event through the spread of radiation via air and water.

The Chernobyl nuclear disaster serves as perhaps the first global crisis in human civilization history triggered by a very small group of individuals. In the years following, the possibility of global crises ignited by single entities or small groups has only increased in likelihood and frequency, extending beyond environmental issues.

This escalation is a consequence of “globalization” transitioning from a concept to reality. On top of the natural media of sea, land, and air, humanity has constructed interconnected networks of trade, finance, and information, further amplifying the spread and impact of risks.

Just as the synergy of global trade, finance, and information networks has brought positive effects to the economic domain, they similarly amplify and synergize risks in the realm of global threats. For example, the mere connection of trade networks was enough to bring about the potential for racial extinction to the indigenous peoples of the Americas. Similarly, the advent of the information network has meant the emergence of new risks from nothing.

The internet has brought the production and distribution of risks to a new peak of decentralization and individualization, something trade and financial networks had not achieved.

When trade networks were established, risks to a country often came from another country, such as the infamous opium trade. With the advent of financial networks, corporations could pose risks to an industry or a country, as seen in the actions of Wall Street investment banks towards Japan and Korea following the Plaza Accord and the 1997 IMF Agreement. In the era of the internet, individuals can impact corporations, industries, or even countries with just a statement or through hacking.

The threshold for becoming a producer of risk is lowering, while the scale and frequency of victimization are increasing.

Taking nuclear leaks as an example, in the internet era, a crisis could potentially be triggered by an individual. All that’s needed is the development of a virus capable of infecting nuclear facilities, and indeed, such actions have already been taken.

Beyond such “high-level threats,” the interconnectedness provided by information networks also exposes us to risks associated with information itself.

For instance, the Blue Whale Challenge, a suicide game invented by a disturbed individual, has led to the deaths of nearly a thousand children worldwide. A major data breach in a large Chinese travel website caused widespread privacy leaks across unrelated industries. The piracy site Manga Village caused significant revenue loss to the entire Japanese manga industry.

In addition to affecting individuals and industries, informational risks can also unexpectedly impact national affairs. Rumors and fake news constantly influence the political ecosystems of countries, leading to uncontrollable consequences. Even the United States, which once viewed social networks as beacons of democracy encouraging their role in social change, was bitten back by a series of fake news incidents during the 2016 elections.

Objectively, the internet acts as a decentralized risk manufacturing machine, turning every risk into a global concern.

From 2012, the FBI began investigating internet child pornography issues through “sting operations.” Between 2012 and 2016, the FBI took control of at least 21 well-known child pornography sites through plea bargains and hacking, uncovering the entire network behind these sites.

Surprisingly, these sites had viewers and contributors from around the world. Since 2012, the FBI has had to share the crime leads it obtained with public security authorities globally (including China) because the issue had far exceeded its capacity to handle.

In the dark market of child pornography, criminals entice or coerce children into inappropriate acts by offering snacks and toys, especially in economically backward areas. These acts are filmed or recorded and then sold on global child pornography sites, with transactions often made through digital cryptocurrencies to evade regulation.

Just as the internet facilitates legitimate transactions, it also acts as a mediator for illegal ones. It could be argued that without the internet, such a systematic, professional criminal industry chain could not exist. The internet has enabled this risk to materialize from nothing, gathering the demands of sexual predators scattered across the globe and directing them toward vulnerable children in economically disadvantaged areas.

Beyond the direct risks associated with the internet, the integration of information networks with trade and finance networks can trigger crises with even greater and more far-reaching impacts. It’s evident that with the inclusion of information networks, the speed of risk transmission multiplies exponentially, far outpacing the spread of economic benefits. This rapid spread of risk makes traditional preventive and remedial measures seem particularly ineffective.

Next, we will explore why the fusion of these “three networks” transitions the concept of a risk society from theory to reality.

4.2 The Habitual Mismatch of Risk

The internet has bestowed upon us capabilities far beyond those of the past, often leading us to overestimate our capacity to shoulder responsibilities well beyond our means. This habitual mismatch of risk has become a primary mode by which the internet age amplifies risks.

Let’s use the recent trend of “e-commerce alleviating poverty” to illustrate how the synergy between information networks and the networks of trade and finance can amplify risks.

In general, farmers in remote mountainous areas tend to have poor liquidity and low asset levels, with their daily economic conditions barely linked to the broader economy.

Despite using the same currency as urban dwellers, these farmers typically spend on necessities within their village or the nearest town. Their primary connection to the macroeconomy is through selling agricultural products to dealers who visit periodically. Due to the lengthy traditional supply chain, the price received by farmers at the point of sale can be as little as one-fifth or one-tenth of the final retail price in cities.

Over the long term, this arrangement, unfair as it may seem to the farmers, creates a buffer of five to ten times the elasticity to absorb risks. This means that events like the 2008 financial crisis or the 2019 US-China trade friction would not significantly affect the income of a farmer reliant on selling produce.

However, the advent of e-commerce has changed this scenario.

E-commerce alleviation of poverty, much like online shopping in general, achieves its results by cutting out middlemen. It increases farmers’ incomes by enabling them to sell their products directly to urban consumers from the point of origin. While this can significantly increase a farmer’s income in the short term, it also embodies the “poverty alleviation” aspect of the initiative.

The hidden crisis is that “eliminating the middleman’s profit margin” also means “eliminating the middleman’s risk distribution.”

Many farmers, still engaged in small-scale farming economies, are directly linked to macroeconomic cycles. While an upturn in the economic cycle can bring significant benefits, the safety of these small farming economies during a downturn is questionable.

To better understand, let’s consider a hypothetical scenario that highlights potential issues:

Imagine a fruit farmer whose family’s income depends on growing fruit trees and selling the produce to visiting dealers (traditional model). Once the produce is sold to the dealer, the farmer’s labor income is secured.

Enter e-commerce poverty alleviation: the farmer decides to join an e-commerce platform’s poverty alleviation program for better earnings.

With the platform’s support, the role of the dealer is diminished, but the functions they performed do not disappear. Fruits need to be transported, higher quality fruits need to be packaged or semi-processed, and promotional pages for the fruits need to be created on the e-commerce platform.

In this process, tasks traditionally handled by dealers, such as transportation, packaging, processing, and marketing, are partly taken on by the e-commerce platform, with much of it falling on the farmer.

For instance, in the transportation step, where dealers used to pick up produce regularly, farmers now hire logistics companies to transport goods from their orchards. Similar shifts occur in subsequent steps. This means the costs previously borne by middlemen are now transferred to farmers. Farmers might have to pay for logistics, packaging, processing, and rent logistics warehouses in major cities.

Given we’re discussing poverty alleviation, it’s reasonable to assume these farmers may not have the cash on hand to cover these intermediary costs before their produce is purchased by consumers.

At this point, “merchant loans” offered by e-commerce platforms might be their best option. Farmers could borrow from inclusive financial institutions at low interest rates during the harvest season to cover the season’s costs, repaying the merchant loan with the sales revenue, and keeping the remainder.

In an upward economic cycle, if the fruit sells well, this all seems fine. But what happens to the farmer if the economic cycle turns downward? Let’s extrapolate:

  1. Without middlemen, unless the farmer has already started selling on the e-commerce platform, they might not realize, “This year’s economy is poor, which will affect city dwellers buying my fruit.”
  2. The only way the farmer might realize “this year’s sales are bad” is if some intermediary costs have already been paid, by which point the produce has been commercialized, with these costs leveraged as described.
  3. If the fruit doesn’t sell, the farmer doesn’t make the expected income, but the initial loans still need to be repaid.
  4. Further, if the farmer can’t repay the loans, numerous small defaults accumulate at the financial institution, creating a vicious cycle with the “economic downturn.”
  5. Additionally, there’s a subtle difference between “selling to a dealer” and “selling directly on an e-commerce platform”: the former involves immediate cash payment, while the latter involves a payment period.

A downturn in the economic cycle could delay payment terms for the farmer on the e-commerce platform, which might delay payments due to operational reasons or even collapse without repaying the farmer. Either scenario is a fatal blow to an already vulnerable economic entity like the farmer.

The “fatal” in this fatal blow is literal because, from an economic standpoint, small-scale farming isn’t as fragile as commonly perceived. Compared to the modern economy, systemic risks in small-scale farming largely stem from natural disasters.

In other words, as long as farmers retain their land and the simplest agricultural tools, they can start over from scratch regardless of the setbacks. The introduction of modern financial tools, especially layered borrowing, could result in farmers losing their land, homestead, and basic production tools, leaving them with no opportunity to recover.

This mismatch of risk capacities is a familiar concept known as “ecosystem anti” or “ecosystem reversal,” exemplified by the situation with China’s LeEco company.

LeEco’s foray into the automobile and smartphone industries in 2015 could be likened to farmers in need of “e-commerce poverty alleviation,” with playing the role of the “e-commerce” platform providing aid. At the tail end of the mobile internet industry’s growth cycle, using the cash flow from internet businesses to support ventures in other sectors was a common strategy. However, LeEco misjudged the remaining duration of the internet industry’s growth cycle and the time needed for a new venture to achieve “primitive accumulation” of capital. Moreover, its overly integrated structure allowed risks to be seamlessly transmitted across the LeEco conglomerate without any intermediaries to absorb the shock.

Among the wave of “internet” car companies established around 2014 to 2016, only Faraday Future faced the worst situation. While other internet car companies also found themselves in awkward positions, they did not bring their associated internet companies down with them. In contrast, Faraday Future nearly caused the collapse of

The key difference lies in the other internet car companies and their affiliated internet companies employing more traditional separation measures, avoiding the pursuit of a “no middleman” approach in business and financial operations. This separation ensured that risks arising in any one business could not spread indefinitely to other related businesses, thereby amplifying the risks.

4.3 The New Cold War

Though it may seem counterproductive or even like a cure worse than the disease, major countries around the world have, over the past few years, established more conservative market regulations and foreign policies. As a result, some argue that we have entered the era of Cold War 2.0.

However, this new Cold War is essentially a war over individuals, with one side being every person living in modern society and the other being society as a whole.

As highlighted in the previous discussion about the challenges faced by Didi, once we fall into the trap of trying to absolutely prevent all risks, it leads us into a state of perpetual anxiety.

On one hand, we live in a world full of unexpected events. The very existence of humanity and life itself is a series of fortuitous accidents in the universe. This means we can never find a way to prevent all risks in such a world.

On the other hand, the internet has heightened our awareness of risks, constantly reminding us that every individual, organization, and country connected through it could potentially be a creator or a transferor of risks.

This has put our society, from individuals to governments, into a state of tension and sensitivity, leading to overreactions, preemptive reactions, and overinterpretations of some normal negative impacts.

Since we can’t actually predict where the next risk will come from, we are left to treat all potential “enemies” as actual threats.

We find ourselves divided over national stances between countries like the US and China, competitive companies, and even differing fanbases of celebrities. While such divisions have always existed, what’s different now is that technology has armed us with deadly weapons against each other.

This bears a resemblance to the “Dark Forest” theory described in the science fiction novel “The Three-Body Problem.”

On an individual level, traditional neighborhood and community trusts have disintegrated. Helping the elderly has become a sign of flaunting wealth, conversations with friends are feared for potential legal repercussions if screenshots are spread, concerns about being imprisoned by one’s employer or harmed by one’s nanny, or worrying that naive statements made online years ago could lead to social ostracization today.

In the industry, leading enterprises and startups alike adopt more aggressive competition tactics out of fear of “dimensionality reduction attacks” and “cross-industry crushes,” engaging in frequent reporting, framing, and negative PR, and postponing corporate social responsibilities due to survival anxieties.

At the national level, since international trade has never been balanced since the Age of Exploration, the commencement of countries taking meticulous issue with this unbalance signals the beginning of the collapse of international trade, which will also mark the start of the disintegration of basic trust among modern international communities.

Thus, the internet’s enemy is itself: the emphasis on equality has led to divisions, an abundance of trust has bred suspicion, and excessive connectivity has resulted in isolation. The internet has become the Ouroboros and the Tower of Babel of our era, reflecting the self-contradictory nature of modern society.

Today, our anxiety about risks and the countermeasures we take are far from reaching their peak, largely because most people have yet to grasp the essence of risk diversification and individualized production.

In 2003, the first complete human genome was published, a project that united top scientists from six countries: the United States, the United Kingdom, the French Republic, the Federal Republic of Germany, Japan, and China. It took 13 years and over $3 billion to complete.

In 2019, Chinese scientist He Jiankui’s publication on human genome editing sparked global panic due to ethical and medical concerns, leading to his punishment.

However, what most people don’t realize is that as early as 2015, you could buy a CRISPR gene editing kit for just $160 on the crowdfunding site Kickstarter. Other necessary tools for editing mammalian embryos’ genes were almost entirely unregulated and easily purchasable.

According to a Russian institution engaged in human gene editing research13, a legitimate human gene edit currently costs only $15,500, with most of the cost going towards essential monitoring and observation. For someone willing to ignore safety and compliance, the cost could drastically reduce.

A similar case happened with the face-swapping app ZAO. After its initial surge in popularity, users quickly questioned whether it might lead to misuse of their faces, such as having their likeness swapped into illegal videos.

However, the public’s anxiety and concerns about ZAO were not only belated but also ineffectual.

The Deepfake technology used by ZAO had been democratized since 2017, with any malicious individual able to find simple, usable code on GitHub. The potential for misuse of ZAO pales in comparison to the possibilities offered by the already open-source Deepfake technology, especially since ZAO is a commercial project operated by a business entity with profit, not crime, as its driving force.

Indeed, companies like Google had already begun developing counter-recognition technologies for forged videos and had a preliminary solution two months before ZAO caused public panic.

These examples illustrate a point: the anxiety brought about by risks serves no purpose other than hindering technological progress and self-repair.

Technology seems to be descending upon humanity in the manner of a horror novel, but fear seems futile. Unless we intrude into every neighbor’s home daily, inspecting every closet and corner, we can never be sure our neighbor isn’t a mad scientist creating zombies.

In “Risk Society,” Ulrich Beck proposed a relatively straightforward approach to address the societal risks brought about by technological progress. He suggested that consensus should focus on progressiveness rather than on technology itself, acknowledging that technological advancements have historically propelled visible societal progress.

However, this consensus is too vague. It might have been feasible in the era “Risk Society” was written, but now the speed of risk or risk perception spread far exceeds the spread of such consensus.

Reid Hoffman, co-founder of LinkedIn, in his book “Blitzscaling,” seeks to provide a rational framework for rapidly expanding companies to manage risks.

This approach doesn’t depart from the previously discussed paradigm of “expanding risk prevention” but finds a practical and effective way to approach risk without exaggerating its importance to the point of stifling competitiveness.

In “Blitzscaling,” risks are categorized into four types based on the actions required:

  1. Take decisive action immediately;
  2. Take short-term action now and permanent action later;
  3. Acknowledge the issue now and commit to action later;
  4. Let it run its course.

The main criteria for judgment are the current stage of the problem and whether it’s a systemic, global issue. The exposure stage of the problem determines whether to act now or later, and its global nature dictates the type of action.

For some unsolvable issues where the benefits outweigh the downsides, letting them burn out and resolve themselves (similar to the concept of alienation in humans) is advised.

Beyond the corporate realm, this framework can also offer insights for individuals and nations. However, the challenge is that we are emotional beings prone to panic, anxiety, depression, and anger, making it hard for individuals, businesses, or nations to always act as rational decision-makers.

This text can’t provide a definitive solution to the new Cold War or the risk society (otherwise, I wouldn’t need to be writing this), but it does offer a less anxious way of thinking:

While we might inevitably enter a new Cold War era, from a longer perspective (the span of a human life), it may not be a bad thing.

During the last Cold War, human aerospace, agriculture, medical, and energy technologies made leaps and bounds, and even the internet technology I currently fret about was born in the Cold War era. Those who endured injustices during the Cold War will receive historical compensation in their later years.

Just as no one can stop history’s ascent, no one can halt its swings. Therefore, we needn’t fret over problems we can’t solve.

Sit back and relax; let’s move towards the conclusion.

5. Epilogue: Believe in Hope, Not Illusions

So far, the tone of this discussion might seem pessimistically skewed. However, I want you to understand that my intent is not merely to convey a sense of doom, nor do I wish to negate the immense benefits that the internet and its technological revolution have brought to the global stage over the past two decades.

No new technology is a dead end; the periodic growing pains are always just the process of technology and society self-correcting and laying the groundwork for the next revolution.

In fact, it’s precisely because the proliferation of internet technology is comparable to the first and second industrial revolutions that we are experiencing the current discomfort.

In 1878, a 31-year-old Edison showcased one of the world’s first incandescent lamps at the third World’s Fair in Paris. In the nearly 150 years that followed, electric light, much like Prometheus’ fire, has illuminated every corner of human civilization. However, in their early days, electric lights, wiring, and power transmission systems all faced significant societal skepticism. It’s understandable since electricity remains as dangerous as fire to this day, and the early electric industry obviously lacked the safety measures that protect its application now.

As with most historical trends, the proliferation and utilization of technology also have their pendulum swings, with each significant technological change followed by a wave of reflection on that technology. Eventually, new technology reaches a balance that benefits human society the most, through a tug-of-war between progress and reflection.

The pain for technology pioneers (like the internet companies and professionals in this round of technological revolution) is undeniable, as their “privileges” will be revoked.

One manifestation of this privilege is the regulatory and legal framework often lagging behind technological innovation, allowing for transformative changes (mostly in terms of business models) beyond the original capabilities of the technology.

Typical examples include ride-sharing versus traditional taxis, short-term rentals versus traditional hotels, and internet finance versus traditional finance.

In these sectors, technological advancement was merely a catalyst for change. The early disruptive impact of ride-sharing, short-term rentals, and internet finance was based on their ability to operate outside the regulatory frameworks of their traditional counterparts, creating more aggressive products.

This privilege of breaking through existing frameworks was initially more beneficial than harmful. With the introduction of new technologies and a somewhat relaxed environment, industries could identify where new technology could bring growth to the sector or benefits to users.

However, as the technological dividend peaks, the rights to break regulatory frameworks gradually become more harmful than beneficial, leading to the revocation of this privilege.

Nowhere is this more apparent than in the financial sector. The emergence of products like Yu’E Bao, breaking through regulatory boundaries, ushered China into the era of inclusive finance, attracting many users to start managing their finances and helping numerous small and medium-sized merchants improve their business with financial tools. However, the continuous negative impacts of campus loans, consumer loans, and P2P lending have gradually led to the revocation of the internet’s privileges in this domain.

Whether it’s the early breakthroughs or the eventual return to regulation, both phases are beneficial to the overall user base.

Of course, for internet practitioners, the revocation of such “privileges” is manifested through more tangible pains like declines in KPIs, business closures, and drops in revenue.

In fact, during the period of privilege revocation, this pain extends beyond industry professionals to affect individuals across the board. From a basic economic standpoint, when companies begin to face regulatory and legal constraints and fulfill their responsibilities, their operating costs inevitably rise and are passed on to consumers—after all, we’re discussing modern enterprises, not charities.

Take ride-sharing as an example. As public demands for safety increase, so do the prices. This price hike should not be criticized because achieving higher safety standards necessitates additional measures by ride-sharing companies, such as more stringent driver and vehicle screenings, regular driver training, paying for high compulsory insurance for operational vehicles, upgrading app security features, and installing vehicle monitoring systems. All these measures come with a cost.

Eventually, we realize that a ride-sharing industry operating healthily doesn’t offer much price advantage over traditional taxis. This poses the question: why did we ever assume that hailing a taxi via an app should be cheaper than hailing one on the street, using the same car and driver?

Yet, ride-sharing is not a “wrong turn” in the history of human mobility. It has significantly reduced the phenomenon of empty rides, making taxis accessible at almost any time and place without the worry of being too remote. It has increased the overall transport capacity, allowing more people to join the transport industry and creating non-professional transport capacities (like carpooling and ride-sharing) to facilitate smoother urban traffic. These benefits, however, are not always visible in the form of individual consumer savings.

The same goes for other sectors. Almost all consumers who previously enjoyed services afforded by technological privileges will experience the pain of these privileges being revoked. For example, we can no longer pay off credit cards for free using Alipay or WeChat, download free songs via search engines, or collect as many discounts from e-commerce and food delivery services.

Thus, during the period of privilege revocation—a relatively short timeframe—we witness a phenomenon: individual welfare seems to decline, while the overall social welfare increases. The costs that were once leveraged for individual benefit are now reallocated to ensure the welfare of the society as a whole.

For internet companies and industry professionals, this is undoubtedly a period of intense pain. On one hand, the revocation of privileges directly impacts their performance; on the other, the perceived decline in individual user welfare leads to a significant drop in short-term reputation, making them feel caught between a rock and a hard place.

In 1878, “electricians” like Edison, Tesla, and Bell were the coolest people in the world. But today, you wouldn’t harbor any illusions about the electrician profession. Decades from now, when the internet is completely demystified, today’s “geeks” will just be one of the thousands of ordinary professions we take for granted.

This inevitability is unescapable. Hence, we should believe in hope, not in illusions.

What, then, is hope? And what is illusion?

Given the “meta-narrative” context of this text, I won’t provide any explicit answers to avoid leading readers into potentially futile endeavors. However, we can certainly debunk some illusions here.

In today’s internet industry, the concepts of blockchain and decentralization represent a typical illusion.

Some tech pioneers, especially those on the cutting edge of technology, portray blockchain and decentralization with a zeal reminiscent of the idealistic visions of a classless society during the era of class struggles.

However, history teaches us that while capitalism has led to numerous social issues over the past two centuries (and continues to do so), it doesn’t mean everything associated with capitalism should be eradicated from the face of the Earth. Moreover, attempts to do so have brought great disasters to human society as a whole.

Over the last half-century, as modern civilization has grown more rational and mature, whether one subscribes to communism or capitalism, we’ve come to understand that it’s unnecessary to strictly follow the original texts of these ideologies in guiding every step of our societal activities.

Thus, socialist countries have gradually recognized the unique and irreplaceable role of the market in the distribution of production materials, and that private ownership, as a motivational distribution mechanism, can efficiently achieve its goals under current conditions. Meanwhile, capitalist countries no longer view labor purely as a commodity, achieving some degree of worker control over production resources through mechanisms like employee stock ownership and high social welfare.

Except for a few regions, humanity globally has abandoned the binary opposition narrative structure in practice.

Meta-narratives continue to serve as a unifying belief system. For instance, we still champion hard work and success, morally condemn those who gain without labor, lean social welfare towards workers, and implement resource distribution systems primarily based on labor contributions. However, we shouldn’t apply a universal piece-rate wage system or work-point system in specific micro-level practices.

The same principle applies to the internet industry. Critiquing centralized internet successes like Google, Facebook, Alibaba, Tencent, and ByteDance does not mean we should completely revolutionize at the micro-level to build an entirely new, decentralized network—since that would deny the welfare that centralized internet services have already provided us.

To argue further, decentralization is not the panacea for current internet issues. The utopian vision of a “cyberspace manifesto” is unattainable unless you can magically conjure up internet infrastructure from thin air.

Take, for example, the idea of “breaking up Facebook” (and similar platforms) – it’s a proposal that’s both misguided and impractical.

Anyone with a clear understanding of our needs—for a single place to find all our friends—would realize that centralized processing is not only more efficient but also safer.

A decentralized system would not only degrade user experience but could also exacerbate privacy concerns.

Centralized security is akin to a fortress (like Facebook), where we need to focus on the walls, the castle, and the guards (Zuckerberg). On the other hand, a decentralized approach would require us to monitor every node within the system, as any one of them could potentially become a source of privacy breaches.

Pursuing a decentralized internet through public pressure, without a viable technical alternative to the current centralized applications, is to repeat the mistakes humanity has made throughout history.

A more sensible approach would be to continuously refine the existing internet model through technological and procedural adjustments to address its real issues, rather than blindly striving to construct a new internet to overthrow the old one.

A case in point is Facebook’s Libra—not that Libra is necessarily a good project in itself.

The significance of Libra lies not in whether it employs decentralized technology but in Facebook’s attempt at a new form of decentralized corporate governance. This could effectively address many problems arising from Facebook’s centralization.

Traditionally, Facebook, with its dominant position in social networking akin to a combination of WeChat, Weibo, and TikTok, could launch a financial payment system built solely by Facebook, with other parties merely participating as users. In this model, Facebook would bear all responsibilities within the payment system, even though sometimes the issues might be caused by the participants, not Facebook itself.

Under the Libra model, Libra members would jointly govern the entire payment system, sharing benefits and risks. The members collectively decide which data can be used in finance and which cannot. Each member has its independent risk assessment system and performs cross-checks within the Libra ecosystem.

This model prevents scenarios where a product manager’s whimsical decision leads to millions of users’ data being authorized to an entity across the ocean, never met before, resulting in the company’s CEO having to apologize to Congress years later for decisions he never made.

Is Libra decentralized? To purists in the decentralization movement, it clearly isn’t. Technically, Libra might even resemble a centralized product more closely. However, it undeniably incorporates elements from decentralization theory, addressing issues previously inherent to centralized systems more efficiently than rebuilding a decentralized payment system might.

We should be thankful that no digital currency aligning perfectly with decentralization purists has achieved true success. Otherwise, we would face an era of constant economic crises and the collapse of traditional monetary systems.

The control and processing rights over Facebook’s 3 billion user data may eventually be liberated from the hands of Mark Zuckerberg or the company itself, but not through Zuckerberg’s resignation or antitrust actions against Facebook. In fact, forcing Zuckerberg to resign or breaking up Facebook would likely lead to more severe data breaches and privacy violations, directly contradicting the actual demands behind public outcry.

Although Libra seems likely to fail in its current form, it represents a beneficial attempt that should be encouraged and multiplied.

This brings us back to the two forms of “extinction” that sparked this lengthy contemplation: the extinction of technological bifurcation and the extinction of technological proliferation.

There are two directions here, already recognized by internet giants in the past two years:

To address the extinction of technological bifurcation, giants like Microsoft, Amazon, Google, and BAT have proposed the concept of the “Industrial Internet,” venturing into previously untouched domains with novel technology combinations, in hopes of reigniting this engine.

To combat the extinction of technological proliferation, giants like Microsoft, Google, IBM, and BAT have embraced the “Tech for Good” initiative.

“Tech for Good” is not just about social responsibility; it’s a competitive strategy in the next wave of the consumer internet. When everyone can develop similar search engines, instant messaging, social networks, and video tools, the company that resolves the negative issues associated with these conveniences gains an edge in the next round of business competition.

This is not merely about tech companies or their leaders “being good” but about finely balancing social and commercial interests in practice—a challenge as complex as any sophisticated product innovation.

Take the issue of excessive data collection, for example. The current public discourse is somewhat misdirected.

Many large companies’ apps tend to over-collect user data and permissions, primarily driven by the need to combat fraud. This is particularly true in promotional and customer acquisition efforts, to prevent exploitation by black-market operators using bulk registration techniques to game the system.

Luckin Coffee’s response last year to criticism from the Ministry of Industry and Information Technology illustrates this point well: the excessive data collection was justified as a measure against fraudulent redemption of coffee coupons.

On the technical level, strategies against fraud are relatively straightforward: collect ample data from users and use social engineering along with machine learning to identify potential duplicate users, thereby increasing the difficulty for fraudsters to create fake accounts. However, this is precisely what regular users disdain.

Users might argue, “I don’t care for promotions or gaming the system, so can you stop collecting my data?” This seems like a reasonable request. However, even as of 2019, marketing activities remain the largest drivers of revenue for internet companies. If a company opts for a blanket “ethical” approach, it risks being outcompeted by those who continue to exploit user data.

To arrive at the right solution, we first need a clear problem statement. Past discussions on data privacy have failed to adequately address the issue of fraud faced by businesses, making it challenging to find genuine solutions.

As the peak of the internet technology revolution passes, a more effective approach than hasty problem-solving is to clearly define the problems. Accurately framing these issues can enhance the competitive edge of business products—imagine search engines without medical ads, social networks free from information noise, and recommendation algorithms that do not invade privacy.

However, neither the Industrial Internet nor the Tech for Good movement is guaranteed to address the fundamental challenges facing the internet industry.

Once a technological revolution enters its latter stages, our goal shifts from preventing a downward trend to making the curve as smooth as possible. This is why Silicon Valley, known for its computing and internet innovations, started shifting focus to new domains like renewable energy and biogenetics years ago.

If fantasy is the belief that “one company, one product, one new technology can solve all our current problems,” then what is hope?

Hope lies in the idea that every “problem” we face today will become a “demand” in the near future, which will then transform into products, solutions, and new drivers for societal development in the next era.

After all, this is how human progress has always unfolded.


  1. Internet/Broadband Fact Sheet, Pew Resarch Center, 2018 ↩︎
  2. Teens, Social Media & Technology, Pew Resarch Center, 2018 ↩︎
  3. American Family Survey, CSED@BYU & Desert News, 2019 ↩︎
  4. Media Use by Tweens and Teens,The Common Sense Census, 2019 ↩︎
  5. Study: Stark differences in media use between minority and white Youth, Northwestern University, 2011 ↩︎
  6. Group Chats Are Making the Internet Fun Again, NYK, 2019 ↩︎
  7. 《中国互联网络发展状况统计报告》,CNNIC , 2019年6月 ↩︎
  8. 《2.5亿精神病患的生存现状:离不开的医院,回不去的家!》, 人民网-健康时报, 2018年7月 ↩︎
  9. 《中国崛起与“元叙事”的终结》, TED, 李世默, 2013 ↩︎
  10. A Genealogy of ‘Globalization’: The Career of a Concept, Paul James & Manfred B. Steger, 2014 ↩︎
  11. How America Ends, The Atlantic, 2019 ↩︎
  12. 《最高法数据显示:传统出租车万人案发率高于网约车》, 北京青年报, 2018年9月 ↩︎
  13. 《“俄版贺建奎”在行动:将编辑人卵子中致聋基因,已有一对夫妇愿意参与》, MIT科技评论, 2019年10月 ↩︎
评论尸 Avatar