*Nexus*: Parts 1, 2, and 3

Meeting Name:

Jagongbi (Let’s Study Liberalism and Soar)

Book Title: Nexus: A Brief History of Information Networks from the Stone Age to AI

Date and Time: Wednesday, May 21, 2025, 9:00 p.m.–9:40 p.m.

Venue: Online via Zoom

Participants: Rudino, J, Born to Run, Jayu, Mori, Nanuri, kdg

Written by: Born to Run

A History of Humanity Seen Through Information Networks, from the Stone Age to AI

In the 1990s, less than 30 years after the concept of “digital” was first seriously introduced to the general public, the inventions of the internet and smartphones rapidly connected people across the constraints of time and space. Within little more than a decade of its emergence, artificial intelligence came to outperform humans even in strategic games such as Go, where humans had long been considered overwhelmingly superior, and, based on its remarkable computational power, it began replacing human functions across a range of fields.

The 21st century. Civilization is changing at an even faster pace. In The Future of Wealth, Alvin Toffler foresaw the end of the industrial age and the arrival of a knowledge-based economy in the early 21st century, but the evolution of artificial intelligence, which is rapidly encroaching on leadership across a wide variety of specialized domains, has placed us on the threshold of a civilizational turning point before we have even had time to respond. This goes beyond mere technological progress and raises fundamental questions not only about human identity but also about the very reason for our existence.

The author argues that the reason we are overwhelmingly more powerful than other species is that we can connect on a massive scale and maintain those connections. At the same time, however, he points out that those very connections are confronting us with unprecedented crises.

1. Human Networks: Documents and Bureaucracy

An idea floating around in one person’s head may be true or may be a delusion. But the moment others empathize with it and agree to it, a connection called “intersubjective reality” is formed. Money, religion, law, and the state are all fictions, yet they are representative examples of intersubjective realities that exert absolute influence over our lives. If a network is a web of connections, and information is the key element that makes those connections possible, then intersubjective reality can be said to be the core information that sustains the large-scale network known as sapiens.

The author explains historical development from the perspective of the evolution of networks. In order to survive the harsh natural environment of the Ice Age, sapiens formed tribes that went beyond small kinship groups through the information contained in “stories,” and with the settlement of agriculture this network expanded further. But there were limits to a system of information transmission dependent on memory, and once people began recording information on clay tablets, a new information class called bureaucracy emerged. Bureaucracy and documents helped networks expand to the scale of the state, making large-scale operation possible.

Networks become more powerful as their connections grow denser. The warp and woof that make up those connections are intersubjective realities, and these are the very source of power. People believe that the beliefs of the community to which they belong are close to the truth, but the central purpose of networks throughout history was not the pursuit of truth but the maintenance of order. Documents regarded as absolute and infallible, such as the Bible or The Communist Manifesto, granted absolute authority to organizations like the Church or the Soviet Communist Party and functioned exceptionally well in preserving order.

The problem, however, is that it is human beings who interpret and apply them. Because humans are beings incapable of avoiding error, all human networks based on intersubjective reality have fundamental limitations. The way these limits could be overcome was by humans acknowledging error and creating self-correcting mechanisms to rectify it. The results were the Scientific Revolution and liberal democracy.

Science approached truth by presenting informational errors or new facts at academic conferences and having them verified by anonymous groups of experts, while democracy made the diffusion of power and mutual checks possible by guaranteeing the existence of independent institutions such as the press and universities. Elections were the final mechanism for verifying power.

But this system, too, has a limitation: the vulnerability of its self-correcting mechanisms. If those mechanisms can be neutralized, citizens can be used as tinder for mob psychology, and elections can degenerate into tools for proving a dictator’s legitimacy. In that way, regression into totalitarianism can happen in an instant. And today, silicon-based information and communication technologies capable of concentrating and processing all information in a single hub serve as new hope for totalitarian regimes and as a grave threat to democracy.

2. Inorganic Networks: Active Tools That Are Always Awake

The 21st century is an age of unprecedented information revolution, and the tool at its center is the computer. If printing in the past accelerated the Scientific Revolution and laid the foundation for modern civilization, the computer is a fundamentally different kind of tool.

A computer is fundamentally an “active” tool capable of making decisions and generating ideas on its own. Clay tablets, radio, and even nuclear weapons were passive tools that operated according to human commands, but computers have already become “agents” that shape society, culture, and history beyond human control. A representative example is the social media algorithm.

In the 2010s, Facebook implemented an algorithm designed to increase user engagement in order to boost advertising revenue. After experimenting on millions of people, the computer concluded that anger maximized engagement and began selectively recommending fake news and hateful content. Thirty years ago in Rwanda, radio was a tool of incitement; ten years ago in Myanmar, algorithms decided on their own to become tools of incitement. The result was the genocide of the Rohingya.

Another feature that distinguishes computers from printing is that they are always on. In the age when humans monitored humans, at least a minimum degree of privacy was guaranteed, but computer networks that operate 24 hours a day can monitor even people’s bedrooms and bathrooms. If technologies like Neuralink, which interpret brain signals, become reality, even human “thoughts” may no longer be free from surveillance.

Is this excessive concern? China is already controlling its citizens through surveillance networks and a social credit system, and Iran is punishing women not wearing hijabs through facial recognition technology. Western democracies, too, are introducing various surveillance systems. In the end, the post-privacy era is inevitable, and always-on computer networks force people to exist in a state of constant connection and constant surveillance.

The most serious danger inherent in inorganic computer networks is the alignment problem. This arises when the goals given to computers do not match genuine human intentions, producing a mismatch in execution. For example, the scenario in which an AI tasked with maximizing paperclip production eliminates humanity and monopolizes resources—Bostrom’s “paperclip thought experiment”—is extreme but symbolically illustrative.

The problem is that because computers solve problems in ways entirely different from humans, people may fail to recognize the danger until serious harm has already occurred. As in the case of Myanmar, when biased human intersubjective realities are transferred into algorithms, computers become unprecedented devices for amplifying error. Human beings were left unable to monitor or correct the network until the tragedy of mass killing had already taken place.

This signifies the emergence of uncontrollable technological power and warns that tools we believe we invented for convenience may create an unknown world designed in ways we cannot understand, within which humans risk being reduced to passive beings.

The author says that the best thing we can do to prepare for such risks is to create institutions capable of immediately identifying and responding when problems arise. This is not the domain of technological solutions, but of political solutions.

3. Computer Politics: Democracy and New Divisions

Through two world wars, liberal democracy proved itself superior to Nazism and communism. But are the information network structures of the 21st century and liberal democracy truly compatible?

For democracy to be sustained, dialogue among participants is essential, and communication technology is indispensable for that dialogue. Democracy can function properly only when both mass media and free communication among individuals are possible. However, 21st-century computer-based information technology threatens democracy in two ways.

First, surveillance technology infringes on citizens’ freedoms and rights.

Second, automation and robotics can bring about mass unemployment.

This combination could repeat a development like the 1930s, when the Nazis rose amid the Great Depression and Germany fell into totalitarianism. Today’s technology is more sophisticated, and democratic systems are more vulnerable.

For example, generative AI such as ChatGPT can already use language more elaborately than humans, and AI bots have extensively infiltrated comment sections and spaces of public opinion. Just as counterfeit currency destroys an economy, if AI bots imitate human conversation and distort public opinion, the very foundation of democracy itself could collapse.

Would totalitarian states such as North Korea or China then welcome these new technologies? Twentieth-century totalitarian regimes sought to centralize all power and information, and this in turn exposed them to the risk of collapse. But in the AI era, centralization may instead become a technical advantage. Surveillance and control through AI help strengthen power.

The problem is that no dictator has ever perfectly controlled the inorganic agent known as AI. Totalitarian power inevitably seeks concentration of power, but the deeper that power becomes dependent on AI, the greater the risk that a human dictator will be stripped of power by AI.

In the end, whether in democracy or totalitarianism, the threat of AI is a problem that transcends borders. We live in an interconnected world, and the failure of one country can immediately become a crisis for all humanity.

In the future, rather than the traditional structure of “democracy vs. totalitarianism,” a new structure of division—“human networks vs. nonhuman networks”—will emerge. And in that situation, no one who hands control over to algorithms will be able to gain any true benefit.

Ultimately, politics comes down to the question of who holds decision-making power.

4. Conclusion

In The Singularity Is Near, Ray Kurzweil called the point at which artificial intelligence surpasses human intelligence the “Singularity” and predicted that after this point humans would evolve beyond biological limitations into upgraded beings. By contrast, Yuval Noah Harari, in Homo Deus, drew attention to the possibility that humanity could become a godlike being through technological progress, but left the ending unresolved. Rather, through the example of how Neanderthals were driven to extinction after losing out to sapiens in competition, he warned that we, too, could disappear at the hands of beings that replace sapiens.

Nexus carries forward that same awareness of the problem while pointing out the phenomenon of computer algorithms gradually seizing the initiative. Rather than focusing on the positive future that advanced science and technology may bring, it presents the possibility of a dystopian development in more concrete terms.

This book asks whether the name sapiens, meaning “wise being,” is truly justified. The history of endless war, discrimination and massacre, and environmental destruction raises a fundamental question as to whether we really are wise beings. And it asks whether we, as we are, can truly handle computer algorithms—far more powerful than any tool we have dealt with until now—“wisely.”

The author says that history is not a straight line toward a predetermined destination, but an open arc, and that the mechanism of correcting one’s own errors is the principle of survival for all living things. Through various historical examples, he argues that self-correcting mechanisms have functioned effectively even within the intersubjective realities of human society, and he believes that this mechanism can also be applied to computer algorithms.

Therefore, before it is too late, we must cooperate to establish mechanisms capable of controlling AI. This is not a technological problem but a political one.

* The “Liberalism Reading Group” Kakao Open Chat and Band group are communication spaces operated by reading group members.

These are not channels operated by the Center for Free Enterprise (CFE), so please keep this in mind when joining and participating.

▶ Join the Liberalism Reading Group:

https://open.kakao.com/o/g4Nn1uud

---

Original title: <넥서스> 제1부, 제2부, 제3부

Author: Jayujuui ipmun dokseotoronmoim

Date: 2025-05-25

Source: https://www.cfe.org/bbs/bbsDetail.php?cid=reading01&pn=1&idx=27895