The House Committee on Homeland Security will host artificial intelligence company Anthropic for a closed-door briefing on Wednesday regarding the company's Mythos cybersecurity model, according to a source familiar with the meeting. The briefing will focus on the model's capabilities and national security implications as lawmakers grapple with the safety risks posed by rapidly advancing AI technology.

The meeting comes amid heightened congressional scrutiny of artificial intelligence systems and their potential security vulnerabilities. Lawmakers have expressed growing concerns about how AI could be exploited by adversaries or create unintended risks to critical infrastructure and national security.

The development reflects broader efforts by major AI companies to engage with federal policymakers. OpenAI and Anthropic have both established Washington offices and hired lobbyists as they seek to influence federal regulation of artificial intelligence technology. The companies are spending unprecedented amounts on lobbying efforts to shape potential AI legislation.

Meanwhile, cybersecurity researchers continue testing AI capabilities in controlled environments. Recent competitions have featured experts and college students using AI agents to attempt breaking into and defending computer networks, with the AI systems showing mixed but notable performance operating independently.

The closed-door nature of Wednesday's briefing suggests sensitive technical details may be discussed that are not suitable for public disclosure. Such private briefings have become increasingly common as lawmakers seek to understand complex AI technologies while companies aim to demonstrate responsible development practices to regulators.