OpenAI CEO Sam Altman says society should accept that “some bad things” will happen as AI advances, arguing the benefits outweigh the harms. In interviews reported by multiple outlets, Altman frames the trade-off around enabling widespread public access and “agency” to use the technology, while stressing that this does not mean accepting unlimited danger.

Several sources report that Altman discusses differences between OpenAI and Anthropic’s approach to risk. He indicates openness to slowing development of the most advanced models and supporting stricter safety requirements and independent evaluation measures, while rejecting the idea that regulators—or a single lab—could guarantee that misuse, hacks, or scams never occur. The reports also note that Altman draws a line against “catastrophic risks,” including scenarios involving loss of control of AI.

Outlets vary in emphasis: some foreground Altman’s willingness to tolerate harms, while others highlight the regulatory and safety context, including OpenAI’s participation in proposals and safety pacts. Overall, the coverage centers on Altman’s argument that some negative outcomes are an expected cost of progress, but not at the expense of the most extreme risks.