What happened
On June 29, 2026, Senator Mark Warner (D-VA) released a draft of the 'Artificial Intelligence Access, Gatekeeper Exchange, and Nondiscriminatory Transfer (AI AGENT) Act' (reported by CyberScoop). The bill would: (1) require large online platforms (50M+ monthly users) to allow users to choose at least one FTC-certified AI agent provider; (2) empower the FTC to create a registry of AI agent software providers certifying their privacy and cybersecurity protections; (3) require all AI agents to be linked to a verified human operator's identity; and (4) mandate built-in controls for users to grant or revoke agent permissions. The bill is currently a draft — not yet introduced as formal legislation.
Why it matters
If enacted, the AI AGENT Act would create the first US federal framework specifically governing AI agent interoperability and identity accountability. The FTC registry mechanism would make AI agent compliance a formal federal certification process. The human-identity-linking requirement would materially reshape how agentic AI systems are architected. As a bipartisan draft from a senior Intelligence Committee member, it signals serious legislative intent and previews the compliance obligations AI agent providers may face.
Action needed
Monitor for formal bill introduction and committee hearings. AI agent developers and platform operators should begin assessing what a human-identity-linking architecture would require technically and legally. Begin engaging with the proposed FTC certification framework in public comments and industry coalitions.