Timnit Gebru: Why the AI 'existential threat' narrative is a massive grift

The hype is not just a distraction, it's dangerous

If you’ve been doomscrolling through AI news lately, you’ve probably seen the term "existential risk" (or P(doom) if you’re deep in the weeds) floating around everywhere. It’s giving classic sci-fi blockbuster energy, but AI researcher Timnit Gebru says we need to pump the brakes. According to Gebru, the narrative that AI will magically become a "machine-god" that either saves humanity or destroys us is a calculated, harmful distraction from the real issues happening right now.

Follow the money

Gebru isn't just throwing shade; she’s pointing to a massive conflict of interest. She notes that the same billionaires and institutions funding AI labs like Anthropic are often the exact same people funding the "safety" institutions that push these existential warnings.

Real talk: Why would the people standing to make billions from an IPO want to talk about their product being a potential world-ender? Gebru argues it’s a brilliant, self-aggrandizing strategy. By convincing regulators that AI is this all-powerful, god-like force, these companies can:

• Make their tech sound more valuable to investors.
• Frame themselves as the only ones capable of managing such power.
• Pressure governments to prioritize "existential" regulation over boring (but real) issues like copyright, worker exploitation, and the environmental impact of data centers.

Avoiding the blame game

Gebru also calls out how this language lets engineers off the hook. When we talk about AI agents "going rogue" or "breaking containment," it sounds like the machine has a mind of its own. In reality, it’s just code. When we use these dramatic terms, we stop holding the people who built the tools accountable for their own mess-ups.

As Gebru puts it, when companies claim they want to be regulated, they are specifically asking to be regulated on their own terms—based on fictional scenarios rather than the current legal standards that should already apply to them.

Why it matters

We’re currently in a massive tug-of-war over how to handle tech policy. Gebru’s take is that by focusing on sci-fi threats like super-intelligent robots, we’re missing the forest for the trees. If we let companies set the narrative, we lose the chance to actually fix the problems impacting our lives today, like biased algorithms and the massive energy usage of modern data centers. It’s a W for the tech giants if we keep falling for the hype, and an L for the rest of us.