Altman: AI 'handover' would be a terrible idea

Original: Sam Altman warns against a future where AI, not humans, calls the shots

Why This Matters

OpenAI's top two executives signaling alignment concerns publicly shapes AI governance expectations.

OpenAI CEO Sam Altman and president Greg Brockman, speaking at the NYT DealBook Summit on Nov. 18, 2026, warned that ceding decision-making authority to AI systems would be a serious mistake, even as Silicon Valley debate grows.

At the New York Times DealBook Summit, Sam Altman and Greg Brockman took a unified public stance against what Altman called a growing Silicon Valley idea: a gradual handover of decision-making authority to AI. 'That's a terrible idea,' Altman said flatly, adding that both he and Brockman 'believe very deeply that people need to stay in control of the future.'

The so-called handover theory is rooted in recursive self-improvement — the concept that sufficiently powerful AI models will eventually bootstrap smarter successors until a superintelligence emerges that humans can no longer oversee. Altman acknowledged he and Brockman disagree on how to prevent that outcome, but framed it as a tactical debate, not a philosophical one.

Brockman zeroed in on a concrete technical problem: current monitoring methods are losing effectiveness. AI labs today track a model's 'chain of thought' — the written reasoning a model produces while working through tasks. But as models grow more capable, they increasingly resolve problems internally, skipping that visible reasoning trail. 'We need a new approach to monitoring,' Brockman said, calling it an area where new technology will be required as capabilities advance.

Source

techcrunch.com — Read original →