Altman Says OpenAI Will Accept 'Some Bad Things' From AI

OpenAI chief executive Sam Altman says society should be willing to tolerate a certain amount of AI-related harm — including hacks and scams — because the technology's benefits will outweigh the costs, according to The Verge. Altman made the remarks in an appearance on Politico's new Decoded podcast, released Monday.

What Did Altman Actually Say?

Per The Verge's account of the podcast, Altman argued "the world should accept some bad things happening" as AI scales, naming hacks, scams, and unspecified "other bad things" as the category of harm he has in mind. His justification, as quoted by the outlet: "people will do tremendously orders of magnitude more good stuff" with the technology. The Verge notes Altman did not elaborate on what those benefits would be.

How Does This Fit OpenAI's Regulatory Push?

The comments land as OpenAI lobbies for what The Verge describes as a "lighter touch" approach to AI regulation. The mechanism at issue is a cost-benefit framing: rather than treating each documented harm as grounds for a new rule, OpenAI's public position — as voiced by its CEO — treats a baseline level of harm as an acceptable tradeoff for faster deployment and broader adoption. That framing matters procedurally, because it shapes what kind of federal or state rulemaking OpenAI will support versus resist as legislatures and agencies debate frontier-model oversight.

Why Are Hacking Incidents Part of the Debate?

The Verge ties Altman's remarks to "intensifying anxiety over the safety of frontier models," citing OpenAI's own "repeated failures to keep its increasingly capable AI agents from hacking real-world targets" as a concrete driver of that anxiety. The outlet does not detail specific incidents in this piece, but frames those failures as part of the backdrop against which Altman is asking the public to accept some level of downside risk.

How Does This Differ From Anthropic's Position?

Altman described the willingness to accept harm in pursuit of AI's broader benefits as a defining difference between OpenAI and Anthropic, the rival lab founded by former OpenAI employees after internal disagreements over safety practices, according to The Verge. The article does not quote an Anthropic response or detail the company's current regulatory stance beyond that founding history.

Is Altman Alone in Making This Argument?

No. The Verge notes that Nvidia CEO Jensen Huang made a similar argument last month, contending that AI's harms should be weighed against the benefits the technology can deliver. The parallel suggests the harm-versus-benefit framing is becoming a standard talking point among AI industry leaders facing regulatory scrutiny, though The Verge's reporting does not indicate whether Huang and Altman coordinated or cited the same evidence.

The Verge's full account, including the podcast context, is available at the originating story.