Yuval Noah Harari: Losing Control of AI Will Be Worse Than Fascism

Humanity could face a future more dangerous than fascism if artificial intelligence evolves beyond human control, historian and bestselling author Yuval Noah Harari warns, urging governments to act before technological power outpaces democratic oversight.

“Fascism is still human. And we have survived it in different parts of the world several times,” Harari said during a discussion with Nobel Peace Prize laureate Maria Ressa and investigative journalist Carole Cadwalladr at London’s Union Chapel on October 5. “The real worst-case scenario is that humans lose control. And once we lose control to something more intelligent than us, we are unlikely to ever get it back.”

The warning comes as investment in artificial intelligence accelerates. U.S. private AI investment reached $285.9 billion in 2025, more than 23 times China’s $12.4 billion, according to Stanford University’s 2026 AI Index. Global corporate AI investment more than doubled in 2025, while organisational AI adoption rose to 88%. Generative AI reached an estimated 53% of the population within three years of its introduction, according to the report.

Harari, author of Sapiens and Homo Deus, said AI represents a break from previous technologies because it can make decisions and act with varying degrees of independence, rather than merely extending human capabilities.

“We haven’t seen anything like it before,” he said. “It’s not like the atom bomb or the printing press, because AIs are agents, not tools.”

“A tool is something in our hands that empowers us,” he said. “AI is something that can make decisions by itself. This is the most important thing to understand.”

Harari does not advocate abandoning AI, which he said could improve medicine, transport and education. Instead, he urged companies and governments to distinguish between deploying existing systems for beneficial purposes and racing to develop superintelligence without reliable safeguards.

“But don’t race to superintelligence,” he said. “The race to superintelligence is different. It is ideological, almost religious. It is unsafe. This is the biggest psychological and social experiment in human history, and no one knows what the consequences will be.”

The risks extend beyond the prospect of machines replacing workers. Stanford’s 2026 AI Index found that AI capabilities continued to advance rapidly, with several models reaching or exceeding human baselines on selected PhD-level science questions, multimodal reasoning and competition mathematics. Documented AI incidents rose to 362 in 2025, from 233 in 2024, highlighting growing concerns about the technology’s effects and the difficulty of managing its risks.

Harari also warned that AI could weaken the information systems on which societies depend. Machines that generate convincing language and adapt messages to individual users could influence political beliefs, public debate and religious interpretation.

“Everything made of words will be taken over by AI,” he said, asking what remains distinctively human when machines can generate language more effectively than people in many contexts.

Ressa, who has reported on online disinformation and political attacks against journalists in the Philippines, argued that manipulation of information can erode public trust and weaken democratic institutions.

“Without facts, you can’t have truth. Without truth, you can’t have trust,” she said. “Without these three, we don’t have a shared reality.”

The 2026 International AI Safety Report, led by AI researcher Yoshua Bengio and produced by more than 100 experts with backing from more than 30 countries and international organisations, found that AI capabilities were improving faster than many experts had anticipated. It also concluded that evidence of several risks had grown substantially, while existing risk-management techniques remained insufficient. The report stressed that experts continue to disagree about the pace of future advances and the severity of potential harms.

Harari said governments still have time to influence AI development, rejecting claims that technological change is inevitable and beyond political control.

“When they tell you there are no different options, this is inevitable, this is deterministic, what they are saying is there is no politics, there is no time for politics,” he said. “And this is still not true today.”

He cautioned, however, that regulation alone may be insufficient if countries lack the computing infrastructure and technical capacity needed to participate in the AI industry. The concentration of investment and advanced infrastructure could leave governments dependent on a small number of technology providers.

The challenge, he argued, is to ensure that the pursuit of technological progress does not undermine humanity’s ability to govern its consequences.

Harari’s closing advice: “Trust the humans more than you trust the AIs.”

Stay ahead of the Stories shaping our world. Subscribe to Impact Newswire and join our 
WhatsApp Channel for updates on global tech, business, and innovation—all in one place.

Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.

Got a story to share? Contact Us to reach a global audience with Impact Newswire.