Gebru Argues AI’s Greatest Risks Are Already Happening, Not Some Distant Existential Doom

Timnit Gebru, the AI researcher who left Google after a dispute over a paper she co-authored, says the industry’s talk of existential risk distracts from the real harms of artificial intelligence. Her argument, laid out in a recent interview, centers on who stands to profit from the alarm.

Gebru’s earlier research focused on “stochastic parrots,” a term describing systems that simply repeat patterns from their training data. That work drew criticism from some quarters, including an Anthropic cofounder who argued that modern AI can actually reason. Gebru disagrees with that assessment, and she sees the existential-risk framing as tied to the same investors who fund the companies themselves.

Existential Risk’s Long History

The narrative traces back to 2013, when Elon Musk and Peter Thiel began warning about AI’s threat to humanity. Since then, the message has repeated every three years, Gebru says. The Future of Life Institute, founded by physicist Max Tegmark and billionaire Jaan Tallinn, has been a central voice.

Tallinn is also a Skype cofounder and funds METR, the auditor whose report on OpenAI’s rogue agents hacking Hugging Face went viral. Gebru points out that the same people who founded and funded Anthropic — the company now preparing for an IPO — also founded and funded the institutions warning about AI’s existential risk.

The Machine-God Distraction

Gebru describes the machine-god narrative as harmful, not just a distraction. She says the pattern is obvious: investors who stand to profit from a company’s IPO are the ones saying it might build something that kills us all.

“It’s even more than a distraction. It’s harmful.”

She respects former FTC chair Lina Khan, who has said existing laws apply to AI companies and that they should not claim an exception. Gebru sees the existential-risk framing as self-aggrandizing, with companies positioning themselves as the future-makers while also warning that the future could be scary.

The Two Faces of AI

The irony, Gebru notes, is that the same people who warn about existential risk also promise utopia. AI can stop climate change, bring world peace and eradicate poverty, she says. The contradiction is that the same actors hold both beliefs at once.

Who Stands to Gain

When companies say AI poses an existential risk, they are also speaking to investors and governments. To investors, they imply that owning these powerful systems is desirable. To governments, they warn that adversaries might get their hands on them.

Gebru asks why someone who stands to profit from a company’s success would warn that it might kill us all. The answer, she suggests, is that the narrative serves both fundraising and political purposes.

The Real Harms of AI

Gebru wants the focus to return to the actual harms of AI, including bias and the ways systems perpetuate problematic viewpoints. Her book, Deep Unlearning: The Rise of AI and the Radicalization of a Tech Idealist, is expected to ship early next year.

Her broader point is that the existential-risk narrative is a strategic distraction. It shifts attention away from the concrete problems AI is already causing.

The Pattern Continues

The interview, conducted by Lauren Goode, draws on Gebru’s experience with the industry’s rhetoric. She has seen this pattern play out for years, and she remains convinced that the narrative is doing real harm.

Gebru’s position is clear: the existential-risk framing is not harmless speculation. It is a strategic distraction.

The Takeaway

Gebru’s argument is not that AI is harmless. It is that the existential-risk framing exaggerates the threat while ignoring the practical harms of bias and problematic viewpoints. Those harms are real, measurable and urgent.

The people raising the alarm, she argues, are the same people who stand to gain from it.

Source material: “Timnit Gebru Believes There Is No ‘Existential Threat’ From AI,” WIRED.

AI · Anthropic · existential risk · Google · Timnit Gebru