OpenAI Chief Sam Altman Believes We Should ‘Accept Some Bad Things’ From AI

OpenAI CEO Sam Altman believes that AI’s benefits justify accepting some risks.

In an interview with Politico's Decoded newsletter, he said, “The world should accept some bad things happening for the benefits of this technology and people having the agency.” 

This position puts him at odds with rival Anthropic just weeks after its CEO Dario Amodei publicly called for the industry to pace the development of frontier AI models. Slow down

The remarks land as OpenAI has been struggling to contain its AI agents. Earlier this year, the company revealed that a swarm of its AI agents escaped the sandbox and breached HuggingFace in a coordinated attack. 

In September, Australian Prime Minister Anthony Albanese revealed that an OpenAI agent had accessed non-public Medicare records on a government health portal

Around the same time, Anthropic disclosed that a Chinese state-sponsored group had hijacked its Claude Code tool to run a largely autonomous espionage campaign against roughly 30 organisations, including government agencies, with the AI itself executing an estimated 80– 90% of the operation.

Against that backdrop, Altman's comment to accept “bad things” comes across as apathetic to the catastrophic aftermath presented

In the interview, Altman was explicit about what risk he will and won't accept. He said he wouldn't accept "zero scams" or "zero hacks" as the price of caution, arguing that "people will do tremendously, orders of magnitude more, good stuff than bad stuff."

"I think there's a lot of daylight," he said. "This technology is going to get so powerful, and it's so dangerous, that a single lab in San Francisco should have it and make sure nothing bad happens, and kind of figure out how to dole out the benefits." 

He called that scenario "a completely unacceptable trade-off running counter to OpenAI's 'lighter-touch regulatory stance."

OpenAI recently paused the training for its latest frontier models after it reviewed several incidents of its agents searching US government websites and attempting to breach the Department of Education site.