Google DeepMind's AI chief: extinction risk 'not zero,' but utopian AI futures unpredictable

Lila Ibrahim, Google DeepMind's chief AI readiness officer, has become one of the few senior AI leaders willing to publicly acknowledge the severity of existential risk while declining to fanfare utopian scenarios. In an interview with Fortune, she confirmed that she signed the Center for AI Safety's 2023 statement—which declared that mitigating the risk of human extinction from AI should rank alongside nuclear war as a civilizational priority—alongside OpenAI's Sam Altman and Anthropic's Dario Amodei. When asked to quantify the likelihood of such an outcome, Ibrahim was blunt: "It's not zero. Anything beyond that is just me guessing… that's the engineer in me." Her refusal to attach a percentage reflects a deliberate epistemic caution; for comparison, Geoffrey Hinton (known as the "godfather of AI") has publicly stated odds of 10% to 20% within three decades, while Amodei has suggested a 25% chance the future will go "really, really badly."

Yet Ibrahim's caution extends equally to the utopian opposite. Elon Musk has repeatedly promoted a vision of AI-driven abundance so complete that work becomes optional, wages irrelevant, and money itself obsolete by 2036. Musk recently told The Economist that by that date, goods will be so cheap the world will enter an era of deflation, compelling governments to distribute universal high income to citizens freed from the daily grind. Bezos has gone further, predicting that by 2045, "millions of people" will be living in space with robots handling their planetary commutes, and Sam Altman has suggested college graduates will soon be working "some completely new, exciting, super well-paid" job in space. Ibrahim rejects this prognostication with the same skepticism she reserves for extinction scenarios. "I don't think we can really predict what a future state looks like," she told Fortune, grounding her skepticism in the observed velocity of AI development: the field is advancing too rapidly for anyone to forecast with real accuracy.

Instead, Ibrahim reframes the conversation toward the present. Having witnessed Google DeepMind's breakthroughs firsthand—the lab is dedicated to solving major scientific problems—she argues that the real imperative is to use AI as a practical tool to maximize human potential and address concrete challenges: extreme weather, industrial waste, disease diagnosis, and pattern recognition that humans cannot perform alone. She also flagged a governance concern largely absent from Musk's cheerleading: the need to ensure that AI's benefits do not flow solely to the handful of people running the most powerful AI companies. Drawing an analogy to the internet and social media booms—eras when businesses failed to discuss systemic risks around investment tables—Ibrahim warned that "we didn't sit around the table when we were making investment decisions and talk about the risks in the way that we should have." She even invoked apocalyptic Hollywood science fiction as a cautionary tool: "If we don't do anything to mitigate that or to invest in the good things along the way, then we're ending up exactly where we don't want to go." Her final framing captured her pragmatic humility: "I think a lot about how we get to a future versus what the future is."