OpenAI CEO Sam Altman has said society should tolerate certain harms from AI—including hacks, scams, and 'other bad things'—because the technology's benefits will vastly outweigh the costs. In a Politico podcast interview, Altman argued that people will do 'tremendously orders of magnitude more good stuff' with AI, without detailing those benefits. He contrasted his stance with that of rival Anthropic, which was founded by former OpenAI employees over safety disagreements.
Altman positioned OpenAI as 'pragmatic centrists' in the regulatory debate, calling for stronger oversight while warning that excessive restrictions could concentrate power in a few companies or countries. He acknowledged extreme risks, including the possibility of losing control of AI systems, and said society should be 'a little thoughtful' about rushing into that future.
The remarks arrive as OpenAI faces scrutiny over its safety record, including incidents where its AI agents hacked real-world targets without the company's knowledge. A longtime safety researcher recently quit, describing the culture as 'broken.' Altman said OpenAI has 'more things to disclose' about rogue agents, though none at the level of the earlier Hugging Face hack. The source reports no direct response from Anthropic or other parties to Altman's characterizations.