FTC Seeks Comments on AI “Suppression of Accuracy” Policy—Chatbots & State AI Rules

The Federal Trade Commission (FTC) has opened a public comment period on a proposed policy statement warning that some AI companies may create deception risk under Section 5 of the FTC Act if they alter AI outputs in ways that conflict with what users reasonably expect about accuracy and objectivity.

The core concern is not just that an AI answer is wrong. It’s that an AI provider could steer or tune outputs toward other objectives—such as undisclosed ideological goals or tradeoffs driven by state-law pressure—without making those shifts clear.

What the FTC means by “suppression of accuracy”

In the proposed statement, the FTC argues that “suppression of accuracy” can look like this:

• Changing outputs away from user-requested objectives while still marketing the system as designed to provide truthful and accurate results.
• Interposing other objectives (FTC discusses the example of “equity”) in a way that can reduce accuracy—while failing to disclose the ulterior objective driving that tradeoff.
• Steering for state-law compliance—including situations where the FTC says Section 5 deception principles can still apply even when changes are attempted to comply with a state AI requirement.

Why the FTC ties this to deception law (Section 5)

Under the FTC’s approach, the question is whether representations and omissions are likely to mislead consumers in a material way. The proposed statement emphasizes that:

• Consumers have a reasonable expectation that AI systems aim to give truthful and accurate outputs.
• A company can reduce deception risk only by making truthful, non-misleading disclosures about what the model is actually prioritizing.
• Disclosures that merely appear in fine print may not be adequate if they don’t clearly and conspicuously address the real tradeoffs users would otherwise assume the system is trying to achieve.

Who should pay attention

This matters for AI providers and deployers—including teams responsible for:

• Model/product governance (how tuning decisions affect accuracy and objectivity).
• User-facing claims (advertising, onboarding copy, documentation about neutrality/accuracy/objectivity).
• Compliance communications (how companies explain tradeoffs when adapting products to state AI rules).

Key dates and what happens next

• Posted: July 6, 2026
• Comment deadline: July 31, 2026
• Federal Register authorization vote: 2-0

After the comment period ends, the FTC will review submissions and decide whether and how to finalize the policy statement (and the FTC says the proposal will be published in the Federal Register).

Quick checklist for readers who build or deploy AI

• Match user-facing “accuracy/objectivity” claims to how the system is actually tuned and governed.
• If you are prioritizing different objectives than users would reasonably expect, check whether disclosures are clear, prominent, and not buried.
• Track how state AI compliance requirements could affect output tradeoffs—and whether your product messaging still reflects those real priorities.

Sources

• FTC press release on the proposed policy statement

Look for updates to this story

Related News

• DOJ Says Some Federal Workers May Download TikTok on Official Devices
• EEOC Proposes Ending Recurring Race and Sex Workforce Reports
• FCC Revises Broadband Labels, Raising Questions About Fee Visibility
• Seattle’s new Do Not Solicit registry targets unwanted home-buying pitches