WASHINGTON — OpenAI chief executive Sam Altman told the Senate Commerce Committee on Wednesday that advanced artificial intelligence systems posed control risks requiring federal oversight, days after warning publicly that humanity could lose control of the technology.

Altman and Anthropic co-founder Dario Amodei appeared before the committee to address safety governance for frontier AI models. Both executives had issued stark statements about existential and economic dangers that drew responses from lawmakers in both parties.

Amodei predicted that powerful AI could displace a substantial share of entry-level white-collar jobs within years. Altman pressed for a federal licensing or testing framework for the most capable models. The two men differed on the pace and severity of near-term risks, with Amodei urging tighter mandatory evaluations before deployment.

Senators from both parties questioned whether voluntary safety commitments were sufficient, noting the absence of binding federal AI legislation. Consumer and labour advocates said the hearing highlighted the gap between industry warnings and enforceable rules.

The session added momentum to competing legislative proposals on AI transparency and pre-deployment testing, though no bill has advanced to a floor vote in either chamber.

"If this technology goes wrong, it can go quite wrong," Altman told the committee, echoing testimony he first gave to the Senate in 2023.