Anthropic chief executive Dario Amodei outlined a three part proposal on September 12, 2026, for what he called pacing the frontier of AI development. The first element would embed third party evaluators, drawn from organizations such as METR, inside AI companies with access comparable to internal risk assessment teams, similar to bank regulators embedded with employees. The second calls for AI companies in democratic countries to agree on common safety standards and limits on the pace of what Amodei called unchecked AI progress. The third envisions international coordination, including attempts to reach understandings with authoritarian governments on narrow, especially dangerous uses of AI. Amodei wrote that because coordinated pacing among competing companies could otherwise raise antitrust concerns, the US government should mediate or at least enable these discussions and issue a narrow waiver for certain kinds of safety conversations, while stressing the government would not need to participate directly. OpenAI chief executive Sam Altman responded that he agreed the embedded evaluator idea was a good idea and said the industry needs to pace the frontier, committing to share more details soon, though he did not separately address the antitrust waiver request in his public response. Antitrust commentators have noted that an agreement among competitors on how fast to develop products is exactly the kind of coordination antitrust law is designed to scrutinize. Days earlier, OpenAI had separately raised the same antitrust question with lawmakers, according to reporting attributed to WIRED, asking members of Congress whether industry wide coordination among AI companies to slow development would violate the Sherman Act, an inquiry that followed a September 6, 2026 blog post by OpenAI chief scientist Jakub Pachocki calling for the AI research community to consider coordinated slowdowns.