Anthropic Co-Founder Confirms the Company Briefed the Trump Administration on Mythos

Anthropic co-founder Jack Clark has publicly confirmed that the company briefed officials in the Trump administration on Mythos, its most capable frontier model, providing the first substantive explanation for why the AI safety-focused company engaged with an administration whose tech and AI policies it has otherwise been challenging through litigation and public advocacy. The confirmation came in response to questions raised by reporting on Anthropic's regulatory positioning.

The Briefing Context

Clark's confirmation addresses what had become an apparent tension in Anthropic's public posture: the company has been vocal about its concerns with the Trump administration's approach to AI safety regulation, has joined industry coalitions opposing certain executive actions, and has maintained public commitments to independent safety standards that place it in opposition to a deregulatory agenda. Against this backdrop, news that Anthropic had conducted government briefings on its frontier model raised questions about the consistency of the company's position. Clark's explanation frames the briefings as a form of safety-motivated transparency — ensuring that government officials have accurate information about frontier model capabilities as they make policy decisions — rather than as lobbying for favorable treatment.

The Dual-Track Strategy

What Clark's confirmation reveals is a deliberate dual-track approach that several frontier AI labs have adopted toward the current administration: maintain independence and challenge specific policies through formal channels while simultaneously ensuring access to government officials on safety-relevant technical matters. The logic is that frontier model capabilities are policy-relevant facts that government decision-makers should understand accurately, and that providing that information is consistent with a safety mission regardless of whether the recipient administration's broader agenda aligns with the lab's values. Whether this framing is genuinely principled or post-hoc rationalization of pragmatic government relations is a question that will follow Anthropic as its policy positioning continues to evolve.

Why Mythos Specifically

The choice of Mythos as the subject of the government briefing is significant. Mythos is Anthropic's most capable publicly acknowledged model, representing the current frontier of the company's research. Briefing government officials specifically on frontier capabilities — as distinct from product announcements or safety frameworks — suggests the briefings were substantive technical engagements rather than introductory meetings. This level of access and detail in government briefings has historically been associated with national security relevance, dual-use risk assessment, or strategic positioning ahead of regulatory action. Anthropic has not disclosed the content of the briefings, so the specific framing provided to administration officials remains unclear, but the decision to brief on frontier capabilities at all represents a meaningful form of government engagement that sits in tension with the company's otherwise adversarial regulatory posture.