Sam Altman asks the Senate to regulate his industry

Sam Altman testified before the Senate Judiciary Subcommittee on Privacy, Technology and the Law, chaired by Senator Richard Blumenthal, in the first of what became a series of congressional hearings on AI that year. He appeared alongside Christina Montgomery, IBM’s chief privacy and trust officer, and Gary Marcus, an NYU professor and prominent AI critic.

Altman’s position was unusual for an executive: he asked lawmakers to regulate his own industry. He proposed a new federal agency that would license any AI effort above a certain capability threshold, monitor compliance with safety standards, and have the power to revoke that license. He also called for mandatory independent audits and safety standards enforced across the industry. Asked whether he would want to run such an agency himself, he declined, saying he loved his current job.

The hearing drew both praise and suspicion. Supporters read it as a sign that a leading AI company was taking catastrophic and near-term risks seriously enough to invite oversight. Critics — including, later, other AI executives and open-source advocates — argued that licensing schemes of the kind Altman described would be easiest for well-resourced incumbents like OpenAI to satisfy and hardest for smaller labs and open-weight developers, potentially entrenching the companies already at the frontier. Marcus, for his part, pushed for a dedicated federal agency plus international coordination, going further than Altman’s proposal.

No legislation followed directly from the hearing. But it set a template that recurred throughout 2023 and beyond: industry leaders appearing before Congress to argue for regulation in the abstract, while the substance of what should be regulated, and by whom, remained unresolved. It also marked OpenAI’s most visible public engagement with US policymakers to date, ahead of the executive order on AI that followed in October.