FTC Targets Undisclosed Ideological Steering in AI Systems

The Federal Trade Commission published a proposed policy statement on July 1 that takes aim at AI companies that configure their systems to pursue ideological objectives without disclosing those design choices to users. The statement warns that such practices may constitute deceptive conduct under Section 5 of the FTC Act.

The proposal stems from Executive Order 14365, which President Trump signed in December 2025. That order directed the FTC to clarify how federal consumer protection law applies to AI models and to address conflicts between federal standards and state-level AI regulations.

Why it matters

This proposal represents the first major federal attempt to regulate AI output manipulation through consumer protection law rather than content moderation frameworks. For AI companies, it creates potential liability for design decisions that prioritize unstated objectives over user expectations—even when those decisions are made to comply with state laws. The statement also signals the administration's intent to preempt state AI regulations through federal consumer protection authority.

The FTC's deception framework

The proposed statement argues that AI companies have represented their systems as designed to produce accurate, faithful outputs within technological constraints. According to the FTC, consumers accept AI outputs without independent verification more than 90 percent of the time, creating reasonable expectations that systems are not secretly pursuing undisclosed goals.

Under the FTC's three-part test, conduct is deceptive if it involves a representation, omission, or practice likely to mislead reasonable consumers in a material way. The statement concludes that training or configuring models to pursue undisclosed ideological objectives—whether voluntarily or to satisfy state law requirements—may meet this standard regardless of motivation.

The FTC distinguishes this intentional steering from AI hallucinations, which result from technological limitations rather than design choices and do not raise Section 5 concerns on their own.

State law preemption

A central feature of the proposed statement is its treatment of state AI regulations. The FTC specifically identifies Colorado's Artificial Intelligence Act as an example of state law that may pressure companies to suppress output accuracy to avoid disparate impact liability. The statement concludes such laws are impliedly preempted when they conflict with Section 5's purpose of protecting consumers from deception.

Disclosure requirements

The FTC acknowledges that companies can avoid liability through clear, conspicuous, and adequate disclosures that their systems prioritize certain objectives over user expectations. However, the statement sets a high bar: disclosures buried in terms of service would not suffice. The more a disclosure contradicts established consumer expectations, the more prominent and persistent it must be.

Practical implications

AI developers face several immediate considerations. Companies marketing systems as accurate or neutral should assess whether actual design and outputs align with those representations. Organizations subject to state AI laws must evaluate whether compliance steps could be characterized as intentional output suppression conflicting with consumer expectations. Companies that do configure systems to prioritize unstated objectives should review whether existing disclosures meet the proposed standard for prominence.

The public comment period on the proposed statement closes July 31, 2026. These details were first reported by Consumer Financial Services Law Monitor.