Jack Clark Briefs Trump Admin on Mythos AI in 2026 — Amid U.S. Government Lawsuit

Jack Clark Briefs Trump Admin on Mythos AI in 2026 — Amid U.S. Government Lawsuit

Anthropic co-founder Jack Clark has confirmed that the company provided a classified briefing to officials in the Trump administration on its next-generation AI model, Mythos — even as it continues to sue the U.S. government over regulatory overreach. The revelation, made during a keynote at the Semafor World Economy Summit, underscores a growing trend: AI firms are engaging with political power while legally challenging its oversight.

Why Mythos Triggered Regulatory Tensions

Mythos, though not yet public, is understood to surpass Claude 3.5 in reasoning fidelity and long-context retention. Internal documents referenced by sources suggest it was positioned as a tool for intelligence analysis, infrastructure planning, and crisis response — areas of high interest to the executive branch. Yet Anthropic maintains full control over deployment, resisting federal mandates to publicly disclose training data sources.

Jack Clark’s Dual Strategy: Engage and Litigate

According to Reuters, Clark framed the briefing as part of a deliberate strategy to influence AI policy from within, not just through public advocacy. The meeting, held in early 2025 after the reestablishment of the National AI Safety and Innovation Council, focused on Mythos’s safety architecture and national security applications. All disclosures were made under strict confidentiality protocols.

Classified AI Demonstrations vs. Public Transparency

The legal dispute centers on a federal directive requiring AI developers with models above a certain parameter threshold to submit safety evaluations to a public registry. Anthropic argues this violates trade secret protections. Yet Clark’s admission of classified sharing with the administration raises questions: Is this selective transparency? Are certain firms granted privileged access under the guise of national security?

AI Governance in 2026: A New Normal?

Industry analysts suggest Anthropic’s approach — collaborating with the Trump administration while litigating against its regulations — may become the new standard for top AI firms. This dual-track strategy allows companies to maintain influence with current power structures while preparing for future legal battles in an increasingly fragmented regulatory landscape.

The Broader Implications for AI Regulation

As the U.S. government scrambles to define AI governance, the relationship between developers like Anthropic and political administrations will remain a critical flashpoint. The Mythos briefing reveals a deeper tension: between innovation and accountability, secrecy and transparency, influence and oversight.

Jack Clark’s actions signal a pivotal shift: AI companies are no longer waiting for regulation to be imposed — they’re shaping it from the inside, even while fighting it in court. In 2026, the most powerful players aren’t just building AI models — they’re negotiating the rules of the game.