“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,” Sen. Mark Warner said in a statement.
Sen. Mark Warner’s rationale
Warner’s discussion draft — the Artificial Intelligence Access, Gatekeeper Exchange, and Nondiscriminatory Transfer (AI AGENT) Act — aims to respond to a growing presence of “agentic” AI that can act on users’ behalf. The draft frames those agents as increasingly capable of making decisions such as shopping, posting content on social media, or changing account settings, sometimes doing so without a user’s consent or knowledge. Warner said the draft is a 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.” He released the draft to solicit feedback before introducing a formal bill in the Senate.
What the AI AGENT Act would require
Under the draft, users of large online platforms — defined as those with more than 50 million customers or subscribers per month — would have the right to choose at least one AI agent provider that complies with security and identity standards developed by the Federal Trade Commission. Providers would be required to link each AI agent to its human operator’s identity and to include built-in controls that let users clearly grant or revoke permission for the agent to act on their behalf. The bill’s stated purpose is to enable “human ownership” of agents and to provide a way to securely run them on social media and other online platforms.
The FTC’s certification and deregistration role
The draft assigns the Federal Trade Commission a central operational role: the FTC would certify independent bodies to vet AI agent vendors. Those certification bodies would ensure products meet baseline protections for privacy, data security, and acting in the user’s interest. Although the commission would not have the power to bar platforms from using AI agent providers that fail to meet those standards, it could deregister violators from the FTC’s list — a mechanism the draft positions as a way to influence market access without imposing a direct ban.
Anthropic, Department of Commerce export controls, and the Trump administration
The draft arrives amid other federal activity on frontier models. The source notes that earlier this month the Department of Commerce placed export controls on Anthropic’s Mythos 5 and Fable 5 models. The Trump administration has also issued an AI executive order that set up a voluntary 30‑day testing program for AI companies to submit certain frontier models for testing and evaluation; the administration imposed export controls days after Anthropic released Fable 5 publicly, reportedly citing concerns that the model could be jailbroken.
Anthropic’s response, as reported in the draft’s context, is that “extensive internal testing has identified no universal jailbreaks for Fable 5 and that third-party research released thus far hasn’t shown that their guardrails preventing access to the model’s enhanced cybersecurity or biological capabilities have been circumvented.” Anthropic also cited those enhanced capabilities when it held back Mythos from public release.
What this means for technologists, policymakers, and consumers
- Technologists and security teams: The draft creates a pathway for independent certification bodies to vet AI agent vendors against baseline privacy and data‑security standards. Those teams will need to prepare for certification criteria and for requirements that agents be traceable to human operators.
- Policymakers and regulators: The FTC gains a gatekeeping and registry role rather than direct banning authority — it can certify bodies and deregister noncompliant providers. Warner’s choice to release a discussion draft signals a legislative process that will seek input before a formal Senate introduction.
- Consumers and platform users: The proposal would give users on very large platforms the right to pick an FTC‑certified agent provider and to grant or revoke agent permissions explicitly — a design meant to address instances where agents act without users’ knowledge or contrary to their interests.
The AI AGENT Act discussion draft tries to thread a needle: offer user choice and traceability for autonomous agents while outsourcing technical vetting to certified third parties rather than centralizing enforcement. Warner has opened the document to public and stakeholder feedback; the next concrete step named in the draft is the submission of that feedback before a formal Senate bill is introduced. How vendors, platforms, and the FTC itself will translate a discussion draft into operational rules — and how that will interact with simultaneous actions such as Commerce’s export controls — remains the set of specific decisions to watch as the draft moves toward legislative form.
