Alphabet board member says US AI regulation is fragmented and confusing

Martin Chavez said the United States is regulating artificial intelligence in a fragmented and confusing way, with regulators judging each new model release on its own instead of setting a clear framework. The vice chairman at Sixth Street and a member of Alphabet’s board made the case in London at Reuters Momentum AI London, a two-day gathering that drew more than 300 vetted CIOs, CDOs, CTOs and their teams.

Chavez argued that the right rules usually follow failure, not before it. He compared the need for AI oversight to the post-crisis financial rules built after earlier market breakdowns, saying the sector needs something closer to bank stress testing: regular, predictable and designed to make systems safer without freezing innovation. His criticism centered on uncertainty, including who is making decisions, what standards are being applied and how companies are supposed to plan ahead.

The policy question has become sharper as the White House has intervened in frontier model releases. OpenAI said on June 26, 2026, that it was delaying a full public launch of GPT-5.6 at the U.S. government’s request, limiting initial access to a small group of vetted partners whose details were shared with authorities. Reuters-linked coverage said the request involved the Office of Science and Technology Policy and the Office of the National Cyber Director. OpenAI said the restricted rollout was not its preferred approach and that broad access should remain the norm. The company’s new lineup included GPT-5.6 Sol, Terra and Luna.

That approach has also shown up in other parts of the AI market. Reuters reported that Anthropic disabled access to some of its most advanced models for users after U.S. government action tied to export controls, another sign that frontier AI oversight is emerging through separate interventions rather than a single published rulebook. The result is a system that can move quickly around national-security concerns while leaving companies to navigate shifting expectations model by model.

The Bank for International Settlements has warned that the AI boom could create its own financial strain. In January 2026, the Basel-based institution said AI investment was surging and could increasingly depend on debt financing, with private credit playing a bigger role. It has also said AI affects output, inflation, labor markets and financial stability, and that the supply chain runs through five layers: hardware, cloud infrastructure, training data, foundation models and applications. Chavez’s remarks captured the trade-off now facing Washington: looser rules may speed deployment, but a patchwork of interventions can also raise compliance costs, unsettle investment and leave consumers exposed.