From the journal

FTC Issues Proposed Policy Statement on AI Accuracy and Output Steering in July 2026

On 1 July 2026, the Federal Trade Commission voted 2-0 to publish a proposed policy statement in the Federal Register describing how AI companies that configure their systems to suppress accuracy or steer outputs toward undisclosed objectives may violate Section 5 of the FTC Act. The statement applies established deceptive-practices doctrine to AI output manipulation. The public comment period closed 31 July 2026.

2 min read

On 1 July 2026, the Federal Trade Commission voted 2-0 to issue a proposed policy statement addressing AI output manipulation. The statement describes the Commission's interpretation of how AI providers that distort their systems' outputs to advance undisclosed ideological or commercial objectives may violate Section 5 of the Federal Trade Commission Act. The statement is proposed, not final, and does not constitute binding regulation; the comment period closed 31 July 2026.

The proposed statement invokes Section 5 of the Federal Trade Commission Act, 15 U.S.C. § 45, which prohibits unfair or deceptive acts or practices in commerce. The Commission's position is that AI providers who represent their systems as objective, accurate, or suitable for a specified task, while internally configuring those systems to suppress accuracy or produce non-neutral outputs, make materially false representations to consumers. The statement does not create a new category of prohibited conduct; it applies established Section 5 deception doctrine to AI output manipulation and steering.

AI platform operators, large language model providers, and enterprise AI vendors that market their products as objective, neutral, or accurate analytical tools must review whether training choices, reinforcement parameters, or content-filtering configurations could constitute output steering under the FTC's framing. Search AI providers and recommendation-engine operators whose products suppress accuracy to favour commercial results are within the policy's stated scope. Healthcare AI companies representing diagnostic AI tools as clinically neutral face similar exposure if output configurations bias results toward undisclosed outcomes.

The statement is proposed and not yet final; the FTC is not issuing enforcement actions under this specific theory pending finalisation. The 2-0 Commission vote reflects a narrower majority than the five-commissioner full bench. The comment record from the July 2026 period will inform the scope of any final statement. Parties that submitted comments may seek to narrow the definition of output steering before the final version is adopted.

Licentium monitors FTC AI enforcement policy and advises AI companies on Section 5 compliance and FTC regulatory positioning. Organisations assessing whether their AI products' output configurations carry regulatory risk under the proposed statement may contact us. Work we undertake includes AI regulatory risk assessment, FTC compliance advisory, AI product disclosure review, and regulatory engagement strategy for AI providers in the US.

Source: Federal Trade Commission, Proposed Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems, 1 July 2026

More from the journal

See all
Illia Prokopiev

EU Instant Euro Payments

Instant euro payments are now a central part of the EU payments framework, bringing faster execution, price parity, payee verification and new access rules for payment service providers. This article examines the binding requirements, key implementation deadlines and practical limits of the regime, while distinguishing the law itself from simplified public descriptions of how the system works.

Illia Prokopiev

South Africa CASP Licensing

Under South African law: When must a crypto-asset service provider obtain authorisation from the Financial Sector Conduct Authority, or FSCA? What category and product authorisations are likely to be required? What transitional exemptions remain relevant? When must the provider register with the Financial Intelligence Centre, or FIC? What ongoing conduct, competence, operational, AML/CFT, payment, and cross-border requirements apply?

MiCA Transitional Period Expires 1 July 2026 as ESMA Orders Unauthorised CASPs to Wind Down

The Markets in Crypto-Assets Regulation transitional arrangement under Article 143 of Regulation (EU) 2023/1114 expired on 1 July 2026. Any entity providing crypto-asset services to EU clients without a MiCA authorisation is now in breach of EU law and must cease operations. ESMA issued a public statement on 23 June 2026 (ESMA75-113276571-1710) setting out expectations for how non-authorised providers must wind down activities and protect investors.