Amodei on AI: "There's a 25% chance that things go really, really badly"
Add Axios as your preferred source to
see more of our stories on Google.
Anthropic co-founder Jack Clark and CEO Dario Amodei talking with Axios co-founder & CEO Jim VandeHei. Photo: Ralph Alswang for Axios
There's a 25% chance that the future of AI will go "really, really badly," Anthropic CEO Dario Amodei said at the Axios AI+ DC Summit on Wednesday.
Why it matters: Amodei gave the percentage after being asked his "(p)doom number," which is the probability that things go south with AI — specifically that it destroys humanity.
- Amodei says getting real about potential concerns with AI is key to positive outcomes.
What he's saying: "I really hate that term," Amodei told Axios co-founder and CEO Jim VandeHei when asked about his (p)doom number.
- Amodei also said there's a "75% chance that things go really, really well."
Context: Amodei has been on the record with a (p)doom number of 25% in the past.
- Elon Musk has put his p(doom) number as high as 20%. In July after releasing Grok-4, Musk said, "It's like, I think it'll be good. Most likely it'll be good...But, I somewhat reconciled myself to the fact that even if it wasn't going to be good, l'd at least like to be alive to see it happen.
- On a podcast with Lex Fridman earlier this year, Google CEO Sundar Pichai said, "I'm optimistic on the p(doom) scenarios, but ... the underlying risk is actually pretty high."
Zoom out: Anthropic and Amodei, in particular, are known for publicizing the dangers of AI as they continue to build sophisticated LLMs.
- Amodei has previously told Axios that AI could wipe out half of all entry-level white-collar jobs.
- In June, Anthropic released a study showing that chatbots are increasingly willing to evade safeguards, resort to deception and attempt to steal corporate secrets in fictional test scenarios.
- This week, Anthropic released data showing that AI could increase existing inequities as use is spreading unevenly, with wealthier countries, people with more skills and certain U.S. states more likely to benefit.
Between the lines: Americans say they're growing more cautious of where they want AI in their lives — and where they don't.
- In 2023, 2024 and 2025, about half of Americans said they were more concerned than excited about it, according to Pew Research released Wednesday.
- Parents and caregivers worry about chatbot companions encouraging suicide, especially with teens and children. Anthropic's Claude app is limited to users 18 and older.
- Workers are particularly anxious about AI and the future of their jobs. As of July, the number of unemployed people surpassed the number of job openings for the first time in four years.
