Google DeepMind chief AI officer signed a statement warning AI could cause human extinction—she says odds are 'not zero' but disagrees with Elon Musk
Fortune Orianna Rosa Royle
Google DeepMind's AI safety chief says human extinction from AI isn't zero odds. She also thinks Musk's utopian 2036 predictions are just guesswork, not fact.
Lila Ibrahim has one of the stranger job titles in tech: chief AI readiness officer at Google DeepMind. Her job, in short, is to worry professionally. And when Fortune asked her point-blank how likely AI is to wipe out humanity, she didn't dodge. "It's not zero," she said, before adding that anything more specific than that is just an engineer guessing.
That quote lands differently because of who said it. Ibrahim signed the 2023 Center for AI Safety statement alongside Sam Altman and Dario Amodei, the one that put existential AI risk in the same sentence as nuclear war and pandemics. Geoffrey Hinton has since put his own number on it: 10 to 20 percent odds of extinction within thirty years. Amodei has said there's a 25 percent chance things go badly wrong. Ibrahim won't hand over a percentage, but she's clear that staying quiet about the risk would be irresponsible, especially with three more years of AI progress behind everyone now.
Her reasoning leans heavily on history repeating itself badly. She points to the pandemic as proof that the disasters people fear aren't always the ones that actually arrive, and she thinks the tech industry has a habit of skipping the risk conversation entirely when the money is flowing. The internet boom didn't include enough boardroom debate about downside scenarios, she argues, and neither did social media's rise. Her fix sounds almost quaint: watch more apocalyptic sci-fi movies and take them as warnings rather than entertainment.
Where Ibrahim gets sharper is in pushing back on the other end of the spectrum, the AI-utopia crowd. Musk recently told The Economist that money will stop mattering by 2036, with AI and robots producing such abundance that wages become pointless and governments hand out universal high income instead. Bezos has predicted millions of people living in space by 2045, commuting via robot rather than train. Altman has talked about graduates landing spectacular jobs in orbit within a decade. Ibrahim isn't buying any of it, and her argument is refreshingly boring by comparison: nobody, including people running the most advanced AI labs on Earth, can actually predict what the future looks like twenty years out.
What she'd rather focus on is far less cinematic — using AI to chip away at extreme weather modeling or industrial waste, and making sure the payoff doesn't just accumulate among the small number of people who own the AI companies. It's a deliberately unglamorous pitch. No Mars colonies, no post-scarcity fantasy, just an engineer trying to keep both catastrophe and hype in check at the same time.
My take
Give Ibrahim credit for refusing to play the prediction game that Musk, Bezos and Altman all seem addicted to — putting hard dates on space jobs or the death of money is marketing dressed up as forecasting, and everyone should be more suspicious of tech billionaires who claim to know exactly how 2036 turns out. The bigger tell here is that the people actually building these systems talk about risk in vague percentages while the people selling the vision talk in specific, confident deadlines; that gap alone should make anyone pause before buying either extreme.
Read more about this at: Fortune
