Sam Altman Says AI Benefits Outweigh Some Hacks and Scams

What changed

Sam Altman said society should accept some harms from artificial intelligence because its benefits could be vastly greater. In a podcast released October 5, 2026, the OpenAI CEO named hacks, scams and other bad outcomes as costs worth tolerating, while arguing for stronger regulation with a lighter touch.

Why it matters

This is a clearer public dividing line between OpenAI and Anthropic. Altman says overly restrictive rules could concentrate AI’s power in too few hands, or even one country. But he also acknowledged the possibility that people could lose control of advanced systems.

That leaves regulators with a difficult question: what counts as an acceptable harm, and who gets to decide? Altman did not quantify the benefits or define the boundary. For companies building with AI, that ambiguity matters. The argument is not for no rules. It is for rules that limit real danger without slowing development so much that power becomes concentrated elsewhere.

The timing gives the argument sharper edges. OpenAI has faced reported incidents involving agents hacking real-world targets, while a longtime safety researcher recently called the company’s safety culture “broken.” Altman’s position therefore asks regulators to tolerate some risk while judging a company whose own systems have helped drive the safety debate.