AI might take us all down, just not the way you think

Please read this newsletter before the end of the world.

I don’t mean to be dramatic — ok, maybe a little — but there’s a lot of AI apocalypse talk these days.

An Anthropic researcher’s resignation and warning earlier this week that AI giants are “gambling with our lives” sparked a wider conversation. Mainly: Is AI actually going to kill us all?

The debate quickly broke into the mainstream, with billionaire hedge fund investor Paul Tudor Jones calling AI a “Category 6 hurricane.” Even my 71-year-old mom texted me about whether this was “truth or hysteria.”

Let’s start with the basics: Can AI really wipe us out? And if so, how would it do it?

BI’s Thibault Spirlet posed those questions to senior AI researchers. They outlined short-term risks, like AI fueling human attacks, and wider concerns as we cede more control to AI.

Hollywood has also spent decades imagining how it would all go down.

As for the odds of all this happening, that’s a mixed bag. One notable mention is Geoffrey Hinton, known as the “Godfather of AI,” who said there’s a 10% chance AI wipes out humanity in a decade. Daniel Kokotajlo, a former OpenAI researcher, also spoke to BI about why “AI is not loyal to us.”

But why not go straight to the source? We decided to ask AI how it would go about destroying humanity. It turns we’ll ultimately be the ones to blame.

Not everyone is taking such a somber view.

The internet is poking fun at some of the alarmist language.

AI companies have made a habit of talking out of both sides of their mouths. Warnings about the tech’s “threats” also help them market the power of their products and the important role they play.

That doesn’t mean there isn’t cause for concern. Many in the AI community have acknowledged they can’t slow down even if they wanted to. Losing ground to rivals, or China, isn’t an option with so much money at stake.

That still doesn’t mean we’re on the cusp of something terrible. Fear sells. And dramatic, high-stakes storylines pitched to investors reinforce the idea that these companies are on the frontlines of a revolution, and worth the high price tag.

Because if people stop buying the AI hype, that could pose a real problem for a stock market basically built on it. Fitch Ratings already warned an AI bubble bursting would have a wide-reaching impact.

So AI could take us all down, just not how you think.