When Dario Amodei published a 5,000-word essay in June 2026 calling on governments to gain statutory power to block AI model deployments, he was making an argument that no sitting AI CEO had put so directly: third-party auditors, not companies, should hold the final veto. Six weeks later, Bloomberg reported that OpenAI's Sam Altman made "many changes" to GPT-5.6 through a "collaborative back and forth" with the Trump administration before its release. The two most closely watched executives in frontier AI have now staked out structurally different positions on the same question, and both are raising money at record valuations while doing so.
The Government Veto Amodei Is Proposing
Amodei's June 2026 essay, "Policy on the AI Exponential," published on his personal site and covered in an exclusive interview with ABC News, argued that AI safety measures are structurally out of step with the pace of capability development. His central proposal: governments should have the power to "block or deter deployment of the model if it is determined, in light of third-party assessment, to present unacceptable risks." The mechanism he described requires mandatory independent audits of frontier AI systems before public release, with the blocking authority triggered by audit findings rather than by any one agency's discretion.
ABC News
A leading news source providing comprehensive coverage across television, radio, digital, and social media platforms.
- Location
- New York, United States
The proposal was accompanied by a $200 million commitment to fund research on AI's societal impact, per Axios. Amodei's framing was explicit about urgency: he described the window between now and 2030 as a period in which AI capabilities could reach what he called a "country of geniuses in a datacenter" level of output, and argued that regulatory institutions need the teeth to act before that inflection, not after. A June 17 CNBC report noted that Washington's actual response went beyond anything Anthropic formally requested, suggesting the essay landed at a moment when policymakers were already looking for a framework.
Critics have characterized the blocking-power proposal as a mechanism that would favor well-capitalized incumbents over smaller entrants, since large labs are better positioned to fund the audits and compliance infrastructure a mandatory review regime implies. Amodei has not publicly addressed the incumbency critique.
Altman's Approach: Negotiate Before You Release
Sam Altman has called global AI regulation "urgently" needed, but his operating practice in 2026 has been consultative rather than statutory. Bloomberg reported on July 9 that Altman described a "collaborative back and forth" with Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent in the weeks before OpenAI released GPT-5.6. The company made "many changes" to the model during those conversations, Altman said. On July 21, Bloomberg reported that Altman is scheduled to brief US officials again on OpenAI's next generation of models, suggesting the bilateral pre-release review has become standard practice between the company and Washington.