AI
Trump Administration Shifts Strategy on Anthropic, Sidelines CEO Amodei
Anthropic cofounder Tom Brown is now leading talks with the White House over exporting restricted AI models, marking a significant shift in the company's engagement with the Trump administration.
WIRED — AI
The Trump administration appears to be taking a softer approach with Anthropic, but with a catch—they're no longer negotiating directly with CEO Dario Amodei. Instead, the company's cofounder Tom Brown and public policy chief Sarah Heck have become the primary contacts in discussions around re-releasing the Claude Fable 5 AI model.
According to sources familiar with the talks, the shift reflects frustration with Amodei's negotiating style. "Tom Brown is not being a weirdo like Dario and can actually engage," one person directly involved in the calls told Wired.
The Export Control Standoff#
The National Security Agency triggered an export ban on Anthropic's most powerful models in June after determining that guardrails protecting the company's Mythos model could be disabled. While the Trump administration hasn't yet lifted those controls, recent discussions suggest there may be a path forward.
The conversations span both high-level strategic talks and technical working groups focused on identifying what proof Anthropic could provide to address the administration's jailbreak concerns.
However, a fundamental problem persists: cybersecurity experts increasingly view AI model guardrails as inherently temporary—skilled users and future AI systems are likely to find workarounds regardless of safeguards. As previously reported by WIRED, this conceptual gap between the company and regulators poses a challenge both sides are still grappling with.
A bipartisan group of lawmakers recently pressed Commerce Secretary Howard Lutnick for clarification on redeployment criteria and timelines, though the Commerce Department has declined to comment on whether it will meet their June 26 deadline for a response.
