Anthropic's Amodei Heads to the White House as Washington Fights Over Mythos Access
April 17, 2026 - 6:32 pm
Summary: Anthropic CEO Dario Amodei is meeting with White House Chief of Staff Susie Wiles on Friday to negotiate access to Mythos, a frontier AI model capable of identifying and exploiting thousands of zero-day vulnerabilities across every major operating system and browser. This follows Anthropic’s blacklisting by the Pentagon after Amodei refused to remove safety restrictions, as multiple US government agencies seek access through Anthropic's controlled Project Glasswing program.
Anthropic CEO Dario Amodei is scheduled to meet with White House Chief of Staff Susie Wiles on Friday in a significant step towards resolving a standoff with the Pentagon over AI model safety restrictions. This comes after various US government entities, including the Treasury Department, intelligence community, and Cybersecurity and Infrastructure Security Agency (CISA), requested access to Mythos, a powerful AI system that poses both remarkable cybersecurity capabilities and ethical dilemmas.
Mythos, revealed on April 7th, is not a traditional cybersecurity tool. It's a general-purpose AI that, during testing, discovered thousands of previously unknown vulnerabilities in operating systems and web browsers. The model achieved success rates exceeding 83% in developing working exploits for these zero-day flaws, many of which had evaded detection for decades.
Anthropic opted not to release Mythos publicly but instead launched Project Glasswing, providing the model to approximately 40 vetted organizations like AWS, Apple, Google, and Microsoft to identify and patch critical software vulnerabilities before malicious actors can exploit them. The company has also pledged up to $100 million in usage credits and $4 million in donations to open-source security initiatives.
The Pentagon Conflict:
This White House meeting is a result of a dispute that escalated since February, when Defense Secretary Pete Hegseth demanded unfettered access to Anthropic's models for various lawful purposes, including potential use in autonomous weapons and domestic surveillance. Amodei rejected this request, leading to the Pentagon blacklisting Anthropic from government contracts with a label usually reserved for foreign adversaries.
Anthropic subsequently filed two federal lawsuits against the Trump administration, alleging illegal retaliation. A federal judge temporarily blocked the blacklist.