Senator Mark Warner (D-VA) released a discussion draft of the AI Agent Act on Monday, the first federal legislation written specifically to regulate autonomous AI agents that act on behalf of consumers.
The bill, whose full name is the Artificial Intelligence Access, Gatekeeper Exchange and Nondiscriminatory Transfer Act, would create three regulatory mechanisms: an FTC registry of trusted AI agents, fiduciary-like duties for agents with access to sensitive user data, and protections for businesses against unauthorized agent access.
What the Bill Requires
Under the draft proposal, AI agents that access a user’s email, e-commerce accounts, or credit cards would be required to “behave in fiduciary-like manner to protect users,” according to the bill’s summary. That means acting in the user’s best interest, protecting their privacy, and safeguarding any data the agent collects.
The FTC registry would serve as a verification layer. Consumers could check whether an agent they’re considering has been vetted before granting it access to sensitive information. Businesses would also gain protections against exploitation by unauthorized agents operating on their platforms.
“As agentic AI transforms how Americans interact with technology, consumers deserve a real choice in the marketplace, and AI agents must be accountable to the people they serve,” Warner said in a statement. He described the draft as “a major step toward building a clear federal framework that promotes innovation, protects consumers and ensures the United States continues to lead the world in emerging technology.”
The Economic Context
The bill arrives as AI agents move from developer tooling into consumer commerce. According to CBS News, Morgan Stanley estimates that AI-powered shopping assistants could account for up to $385 billion of U.S. e-commerce sales by 2030.
Warner is releasing the draft to solicit feedback from stakeholders and the public before formally introducing it, signaling that the final text may change significantly.
Regulatory Timing
The AI Agent Act lands during a broader regulatory scramble around autonomous systems. China’s State Administration for Market Regulation released its own national standard for AI agent identity management last week. The Legal Context Protocol coalition, including Google, IBM, and UiPath, launched an open-source framework for agent transactions on June 29. Warner’s bill is the U.S. federal government’s first attempt to create comparable infrastructure through legislation rather than industry self-regulation.
The distinction between this proposal and existing AI regulation (like Trump’s June executive order on AI model review) is specificity: the AI Agent Act targets agents as a distinct category of software, separate from the models that power them. That distinction matters because agents access user accounts, spend money, and make decisions autonomously, creating liability questions that general-purpose AI regulation does not address.