Senate Proposes Federal Framework for Trustworthy AI Agent Providers

Summary

Senator Mark Warner has released a discussion draft of the AI AGENT Act (Artificial Intelligence Access, Gatekeeper Exchange, and Nondiscriminatory Transfer Act), which would establish a federally vetted list of AI agent software providers that comply with security and identity standards. The bill aims to give users of large platforms (50M+ monthly users) the right to choose from certified AI agent providers while requiring the Federal Trade Commission to oversee certification bodies that ensure baseline protections for privacy, data security, and acting in users' interests.

Under the proposed framework, the FTC would certify independent bodies to vet AI agent vendors, ensuring they link each agent to its human operator's identity and include built-in controls for user consent and permission revocation. Violators can be deregistered from the FTC list, though the bill doesn't bar platforms from using non-compliant providers. Warner framed the bill as addressing growing concerns that AI agents often make decisions on behalf of users—shopping, posting to social media, or changing account settings—sometimes without explicit consent.

The policy context is urgent: Morgan Stanley estimates that nearly 23% of Americans made purchases using AI in a 30-day period last year, with agentic commerce potentially reaching hundreds of billions by 2030. However, AI agents remain unreliable, sometimes making absurd purchases, leaking sensitive data, or acting contrary to user interests. As AI bots proliferate and increasingly interact with one another, the need for accountable human identity verification and baseline security protections becomes more critical.

• The bill responds to rapid AI agent adoption in commerce and social media, with 23% of Americans using AI for purchases and projected growth to hundreds of billions by 2030