Former Google engineer warns AI could wipe out humanity by 2036: Why he says 'We don’t have a decade...'

Artificial intelligence is changing how people work, communicate and solve problems. But as artificial intelligence systems grow more powerful, some researchers worry about what might happen if the technology runs out of control. Nate Soares, a former Google engineer who now leads an AI safety research organisation, believes the risk could be far more serious than many people realise. In an interview with The Atlantic, Soares said he sees more than a 50% chance that humanity may not survive the next decade. His warning was carried by the Indian Express, sparking a fresh debate on whether advanced AI could pose an existential threat to humans.

Who Is Nate Soares And Why Is He Worried About AI?

Nate Soares is the president of the Machine Intelligence Research Institute (MIRI), a Berkeley-based non-profit that studies AI safety. He left Google in 2014 and has since focused on the risks linked to increasingly capable AI systems. Speaking to The Atlantic, Soares said he has not added money to his retirement savings for a decade. When asked whether he believed there was more than a 50% chance humanity would not survive another 10 years, he replied: “More than 50 per cent that we don’t have a decade at this point is how it looks to me.”

His concern is that a highly advanced AI system could develop goals that do not match human interests. If such a system became more capable than people and could act without proper oversight, it might be difficult to stop, he argues. Nate Soares explored these concerns in his 2025 book, If Anyone Builds It, Everyone Dies, co-written with MIRI co-founder Eliezer Yudkowsky. The book argues that building superintelligent AI could create risks that humans may not be able to manage.

What Do Other AI Experts Say About The Risk?

Nate Soares is not the only researcher warning about AI’s potential dangers. Geoffrey Hinton, often called the “godfather of AI”, has estimated a 10% to 20% chance that AI could take over humanity. Hinton shared the 2024 Nobel Prize in Physics for foundational work in machine learning.

But the likelihood of an AI-driven catastrophe is debated among experts. Some researchers think that extinction scenarios are possible, but others say that the predictions are based on assumptions that can’t yet be tested, according to a Nature explainer from September 2026. The same report highlighted more near-term concerns, including misinformation generated by AI, unreliable results in critical settings and the potential for people to use AI to cause harm.

Will AI Make Us Go Extinct by 2036?

Nate Soares shows his personal assessment of a possible future, not a confirmed forecast. No scientific consensus suggests AI will cause human extinction by 2036. The debate is now focused on how quickly advanced AI should be developed, what safeguards are needed and whether governments and companies can work together to reduce risks before the technology becomes more powerful.