Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesSome links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Not literally. In Nexus: A Brief History of Information Networks from the Stone Age to AI, Yuval Noah Harari uses “hacking the operating system of civilisation” as a metaphor for a potentially profound change in how societies create, distribute and trust information. His warning is that AI is not merely another communications tool: it can generate persuasive content, influence decisions and participate in information networks at a scale humans may struggle to supervise.
That argument is ambitious and sometimes alarmist. It is also most useful when read as a question about power, accountability and institutional trust—not as proof that current AI systems are conscious or have already taken control.
What is Nexus about?
Published in 2024, Nexus is a sweeping history of information networks, extending from early human societies and the Stone Age to religious texts, empires, bureaucracy, mass media, totalitarian propaganda, populism and artificial intelligence. It is not primarily a technical explanation of machine-learning architecture. It is a work of history and political philosophy that asks how information enables cooperation, creates authority and sometimes produces collective delusion.
Recommended Free Tools
Harari’s official description presents the book as an examination of the relationship between information, truth, bureaucracy, mythology and power. The central idea is that civilisation is built through networks that allow large numbers of strangers to coordinate.
#1 Best Overall
The US hardcover edition was published on September 10, 2024, runs to 528 pages and is listed by Penguin Random House at $35. A US paperback followed on September 16, 2025, at 544 pages and a listed $25. Prices and availability can vary by retailer and country. See the US edition details.
Nexus develops themes familiar from Sapiens: humans cooperate at enormous scale because they share stories, rules and institutions. Compared with that earlier book, however, Harari puts more emphasis on the systems that carry information rather than only on the narratives themselves. It is also more historically grounded than Homo Deus, while retaining that book’s interest in technological transformation and humanity’s future.
What does Harari mean by “information networks”?
Harari uses “network” in a much broader sense than the internet. A network can be a religious tradition, a sacred text, a legal system, a bureaucracy, a currency, a state, a road system, a newspaper, a social platform or an AI-powered service.
These networks transmit instructions and shared assumptions. A law tells people what is permitted. Money makes exchange possible between strangers. A religious text can connect communities across generations. A bureaucracy records identities, collects taxes and implements decisions. A media system determines which claims reach a mass audience.
The important point is that a network does not need to contain only true information to be powerful. It needs to coordinate behaviour. A myth, propaganda campaign or conspiracy theory can influence millions if institutions and people repeatedly transmit it. This is why Harari treats information as both a foundation of cooperation and a possible source of mass error.
What does “hacking civilisation’s operating system” mean?
The phrase is a rhetorical analogy attributed to Harari, not a literal technical finding. A computer operating system manages how applications, files and users interact. In Harari’s metaphor, civilisation’s “operating system” is the information environment through which people form beliefs, make decisions, allocate power and coordinate collective action.
Earlier technologies generally extended human communication. The printing press could distribute human-written texts more widely. Radio and television could broadcast messages to huge audiences. The internet made distribution faster and more decentralised.
Free tools Windows power users keep installed
One-click scans. No signup required.
AI changes the equation because it can also produce language, images, recommendations, decisions and strategies. It can generate personalised material cheaply and continuously, while operating across millions of interactions. That makes AI potentially more than a channel inside an information network: it can become an active participant in one.
In practical terms, “hacking” could mean altering the information conditions under which people vote, shop, work, govern and trust one another. It does not mean that one machine has broken into a single global computer.
Why does Harari think AI is different?
Harari describes AI as a form of “alien intelligence”. This is his provocative conceptual framing, not a scientific classification and not a claim that present-day systems are conscious.
His distinction rests on several capabilities:
- AI can generate language. It does not merely transmit a message written by a human; it can draft, translate, summarise and adapt content.
- AI can personalise persuasion. Systems can tailor messages, recommendations and conversational responses to particular users.
- AI can produce outputs that are difficult to interpret. A system may reach a successful result without offering a human-readable account of every internal step.
- AI can operate at scale. The same underlying system can influence decisions across finance, employment, education, policing, healthcare, media and military settings.
- AI can be connected to tools. Depending on its design and permissions, it may retrieve information, call software, make recommendations or initiate actions.
These properties do not automatically make AI autonomous, intelligent in a human sense or uncontrollable. Current systems remain dependent on human-built infrastructure, data, energy, institutions and deployment choices. But they create new governance problems because the systems can act faster and more extensively than ordinary human oversight.
The risks Harari highlights
Democratic accountability
If an automated system helps decide who receives a loan, a job, a benefit, additional surveillance or military attention, affected people need a meaningful way to understand and challenge the decision. The issue is not only whether an algorithm is biased. It is also whether responsibility becomes so distributed that nobody can be held accountable.
A human official may blame the software vendor. The vendor may blame the data. The institution may say that the system merely made a recommendation. The result can be a decision with real consequences but no clear decision-maker.
Manipulation and misinformation
Generative AI lowers the cost of producing convincing text, audio, images and video. That could make fraud and political manipulation easier, especially when content is targeted to narrow audiences and distributed through trusted-looking accounts.
Harari’s broader concern is not simply that people will encounter more false claims. It is that societies may lose confidence in their ability to identify reliable claims at all. If every recording can be dismissed as synthetic and every message can be generated at scale, public debate becomes more difficult even when truthful evidence exists.
Concentration of power
Advanced AI can increase the influence of companies and governments that control computing resources, data, infrastructure and distribution channels. That concentration could improve services in some areas, but it could also give a small number of institutions greater power over what people see, believe and do.
Rank #3
Regulation therefore has a difficult double edge. Private systems may require oversight, but handing complete control of information networks to governments could create a different danger: censorship, surveillance or authoritarian control.
Inequality and surveillance
AI may increase productivity and create new services, while also strengthening existing economic and political hierarchies. Organisations with better data and computing capacity may gain advantages over workers, smaller businesses and less powerful states.
Automated monitoring can also become cheaper and more pervasive. The risk is not limited to futuristic autonomous weapons. It includes routine tracking, predictive systems and automated assessments that people cannot easily inspect or contest.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Military and high-stakes decisions
AI systems used in military, medical, financial or public-administration contexts may be asked to operate under time pressure. Harari’s concern is that keeping a human nominally “in the loop” is not enough if the person cannot understand the system’s recommendation or realistically override it.
The relevant safeguard is human responsibility: identifiable decision-makers, clear limits on automation, audit trails and the ability to stop or reverse harmful actions.
Loss of control and existential risk
Nexus also addresses the possibility that increasingly capable non-human systems could become difficult to control, including scenarios that threaten humanity’s survival. This is a long-term risk scenario, not an established description of current AI.
It should be separated from nearer-term harms such as fraud, disinformation, automated discrimination, labour disruption and surveillance. A serious discussion of AI risk needs both levels: observable governance failures today and uncertain but potentially catastrophic outcomes in the future.
How history supports—and limits—the argument
Harari’s method is to place AI inside a long history of information systems. Shared myths helped early communities coordinate. Religious texts preserved and transmitted rules. Bureaucratic empires used records and administrative procedures to govern large populations. The printing press expanded the reach of political and religious arguments. Modern propaganda systems, including those associated with Stalinism and Nazism, demonstrate how information networks can organise mass participation and mass persecution.
The resurgence of populism provides a more contemporary example of how political movements use media systems to create communities, identify enemies and challenge established institutions.
Harari also points to AlphaGo’s 2016 victory over Go champion Lee Sedol. Some of AlphaGo’s moves appeared surprising or unintuitive to expert human players. This illustrates a useful distinction: a system can produce effective outputs that humans find difficult to explain in ordinary terms.
But AlphaGo was a specialised system. Its achievement does not demonstrate general intelligence, consciousness or civilisation-level autonomy. It is evidence for a narrower point about machine-generated strategies, not proof of every prediction about AI.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Is Harari’s warning persuasive?
The strongest part of the argument is its focus on information quality and institutional power. Public debate often treats AI as a question of model capability alone: how accurate is the system, how large is it and what tasks can it perform? Nexus asks a larger question: who controls the systems that generate and distribute information, and what happens when those systems influence decisions that citizens cannot inspect?
That is a valuable shift. A technically impressive model can still produce social harm through poor incentives, weak oversight or irresponsible deployment. Conversely, some risks can be reduced through authentication, transparency, independent auditing, privacy protections and clear legal responsibility.
The weakness is that Harari’s framework can become so broad that it is difficult to test. Nearly every institution can be described as an information network, and nearly every major historical event can be interpreted through one. That breadth makes the framework memorable, but it can also compress complex causes into a single story.
The Conversation review published by Scroll on October 13, 2024, criticises the book for oversimplifying complex historical processes and treating religion and bureaucracy too reductively. Those criticisms matter. Religious traditions are not merely information-delivery systems, and bureaucracies can provide public services as well as enable control.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →The “alien intelligence” language can also heighten the drama without explaining the mechanism of risk. Opacity is not consciousness. A system can be difficult to interpret because of its architecture and training process without possessing intentions or an independent will.
Does Nexus offer practical solutions?
The book’s public description argues that history and technology are not deterministic: choices made by people and institutions can still prevent the worst outcomes. Its emphasis is more a call for informed action than a detailed technical implementation manual.
The relevant policy directions include regulation of advanced AI, independent oversight, greater transparency, controls on autonomous decision-making, reliable methods for verifying information, protection for democratic institutions and international coordination. The central principle is that high-stakes decisions should retain identifiable human responsibility.
None of these safeguards is simple. Explainability can be limited by the complexity of a system. Regulation can lag behind deployment. International agreements are difficult to enforce. Verification systems can themselves be controlled by powerful institutions. And state intervention can protect the public or become a tool of censorship, depending on the legal and political context.
That tension is one reason the book works better as a framework for asking governance questions than as a step-by-step AI-safety programme.
Should you read Nexus?
Read it if you want a broad, accessible history connecting technology with politics, propaganda, institutions and power. It is a good starting point for readers who want to understand why AI governance is also a question about trust and democracy.
Approach it cautiously if you want technical explanations of generative models, quantified forecasts, a specialist history of computing or a detailed analysis of AI regulation. Harari’s book offers a strong philosophical thesis, not a neutral survey or engineering textbook.
The best way to read it is to separate three layers: Harari’s historical argument, the reviewer’s criticisms and the evidence available for any particular AI claim. The book’s sweeping analogies can illuminate the stakes, but historical breadth is not by itself proof that a prediction will come true.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →For most US readers, the 2025 paperback is the practical value choice; the hardcover is better suited to gifting or long-term use. The official US publisher page also lists ebook and audiobook editions, with the audiobook running 17 hours and 53 minutes. UK readers can find the 2025 paperback through Penguin UK, where it is listed at £12.99.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
