The Federal Trade Commission, the US consumer protection regulator, has opened a public comment window on a draft policy that would treat AI tuning for 'ideological' goals as a form of federal deception and preempt state laws like Colorado's AI Act. Comments close July 31, 2026.
The Federal Trade Commission wants the public to help draw the line between an honest AI answer and a federal deception, and the comment window closes July 31, 2026.
The proposed policy statement frames AI companies that tune their systems toward "undisclosed ideological objectives" as potential violators of Section 5 of the FTC Act, the consumer-protection statute that already polices unfair or deceptive business practices. "Deceptive" here is the FTC's working term, not a settled legal category, and the Commission is explicitly asking businesses, developers, and consumers to argue for or against it before the policy is finalized.
In plain English, the FTC is signaling that an AI vendor who markets a model as "objective" or "accurate" but then biases it toward a political, cultural, or partisan view could face an enforcement action for misleading the people who use it. The proposed statement text describes such distortion as at odds with the "explicit and implicit representations" companies make about their systems. The mechanism is not a new rule, which would require notice-and-comment rulemaking. It is the FTC's interpretation of the existing Section 5 standard applied to a new product category.
That distinction matters because the same statement is positioned to do a second job: preempt state AI laws. The FTC's press release names Colorado's Artificial Intelligence Act as a law that "appears to coerce companies into altering the output of their AI models" and says such state laws are "impliedly preempted to the extent [they] conflict with a federal regulatory scheme." The Commission vote authorizing the Federal Register notice was 2-0, and the comment docket is now open on Regulations.gov.
The preemption push traces to a December 2025 White House executive order that directed the FTC to issue a policy statement on state laws requiring alterations to "truthful outputs of AI models." That order, and now this statement, treat federal AI policy as a single national lane and treat state bias-mitigation rules as obstacles to it. Chairman Andrew N. Ferguson tied the initiative explicitly to President Donald Trump's stated goal of "expanding America's global dominance in artificial intelligence."
Legal commentators are already pushing back on the scope. Tech Policy Press argues that the FTC's preemption authority is narrow and that a policy statement cannot displace state consumer-protection statutes the way a federal statute or formal rule might. TechFreedom's analysis reaches a similar conclusion: a policy statement is not a substitute for rulemaking, and any preemption fight will likely land in court. Both pieces treat "ideological" as a contested category whose definition will shape which companies and which outputs end up in the FTC's crosshairs.
The comment window is the structural moment where that contested definition can be narrowed or hardened. Anyone whose work, product, or users sit downstream of an AI system has skin in the game: a developer choosing a default model, a state agency trying to police algorithmic bias, or a consumer whose chatbot is now a research tool. Comments are filed through Regulations.gov under docket FTC-2026-0859 and will be posted publicly once processed.
After July 31, three things to watch: whether the Commission finalizes the statement with the "ideological distortion" framing intact, whether Colorado or another state sues to defend its AI law, and whether the first FTC enforcement actions under the new framing target named companies or stay generic. Each one will determine whether this proposed statement becomes a working federal definition of AI honesty or a political opening that the courts eventually have to close.