• Anthropic CEO Dario Amodei estimates a 25% chance AI leads to catastrophe
• He still believes AI is worth investing in and that the benefits outweigh the risks
• His comments fit with growing public and policy conversations about AI risks and regulation

One-in-four odds might seem pretty good in some circumstances. It’s way better odds than most casino games, for instance. However, it’s apparently unlikely enough for Anthropic CEO Dario Amodei to seem unconcerned after pegging the chance that artificial intelligence leads to a society-ending disaster at 25%.

“I think there’s a 25% chance that things go really, really badly,” Amodei blithely said at the Axios AI + DC Summit when asked about his (p)doom – probability of doom – belief around AI. But he’s more focused on the “75% chance that things go really, really well.”

By “really, really badly,” he doesn’t mean your phone autocorrecting “duck” to something worse. He means scenarios large enough to threaten societal systems, existential risks, badly misused AI, and runaway outcomes that could be catastrophic.

For an industry often drenched in utopian promise or reduced to sci-fi fearmongering, Amodei’s attitude about both the odds of an apocalypse and why he still is pushing forward with the technology did stand out.

Amodei isn’t alone in expressing unease, but he’s in a rarified position. As CEO of the company behind Claude, he’s not a passive observer. He’s shaping the trajectory of this technology in real time. His team is building the very systems whose potential and peril he’s weighing.

If someone told you there was a 1-in-4 chance your car would explode every time you turned the key, you might start walking more. Amodei would apparently become a mechanic and check the car out first before getting in.

AI doom

This also isn’t the only warning issued by Amodei about AI. He’s warned before that AI could eliminate half of all entry-level white-collar jobs and sounded the alarm against U.S. exports of high-end chips to China. That’s what makes Amodei’s framing so useful. It acknowledges the risk, quantifies the uncertainty, but leaves room for agency.

On the flip side, Amodei’s “75% chance things go really well” isn’t optimism for its own sake. It implies the belief that AI could produce enormous benefits for everyone. It could lead to improved medicine, more efficient manufacturing, and perhaps even strategies to address existential crises like climate change (though one key element to solving this might be the energy required for AI models to run).

But the 25% risk demands that those benefits be built carefully, with consideration for safety measures and regulation. Because if the future is 75% brilliant and 25% broken, the question is: What are we going to do to keep the weight on the right side?

You might also like

• Anthropic will nuke your attempt to use AI to build a nuke
• Meta’s AI chief is right to call AI fearmongering ‘BS’ but not for the reason he thinks
• Calling AI chatbots ‘Clankers’ is clunky and clueless

Eric Hal Schwartz is a freelance writer for TechRadar with more than 15 years of experience covering the intersection of the world and technology. For the last five years, he served as head writer for Voicebot.ai and was on the leading edge of reporting on generative AI and large language models. He’s since become an expert on the products of generative AI models, such as OpenAI’s ChatGPT, Anthropic’s Claude, Google Gemini, and every other synthetic media tool. His experience runs the gamut of media, including print, digital, broadcast, and live events. Now, he’s continuing to tell the stories people want and need to hear about the rapidly evolving AI space and its impact on their lives. Eric is based in New York City.