FTC AI Proposal Would Increase Bias and Decrease Accuracy

August 4, 2026 — Press Release

Proposed Policy Would Create Confusion and Reduce Compliance with Antidiscrimination Laws

WASHINGTON – Last week, the National Consumer Law Center (NCLC), on behalf of its low-income clients, submitted comments urging the Federal Trade Commission (FTC) to abandon its proposed policy statement concerning the Suppression of Accuracy in Artificial Intelligence Systems. In its statement, the FTC falsely claims that any attempts to address biases and discrimination in an AI system create “ideological bias” and make the system less accurate, going so far as to claim that removing biases deceives consumers. The FTC’s claims fly in the face of more than a decade of work studying and evaluating AI systems and they perpetuate the widely acknowledged problems with inaccuracies in AI.

“Testing and evaluation of AI systems, including removing bias, make systems more accurate, not less,” said Odette Williamson, director of racial justice advocacy at NCLC. “The FTC’s proposed policy statement creates confusion about how to legally assess bias and discrimination in AI, and it undermines states’ ability to protect consumers from discriminatory AI systems. It should be withdrawn in its entirety.” 

Depending on numerous design and training decisions, AI systems can be infected with bias that hurts consumers. For example, many AI systems are trained on data that reflects historical discrimination. Using AI trained on those old discriminatory models to make consequential decisions regarding access to credit, housing, employment, health care, and benefits can replicate historical discrimination and inflict serious harm on people. Poor design and deployment of AI systems amplifies discriminatory patterns and increases costs to consumers by creating barriers to access in the credit and housing market.

The inaccuracies of AI-generated content are so widely known that Merriam-Webster named “slop,” or “digital content of low quality that is produced usually in quantity by means of artificial intelligence”– its 2025 Word of the Year. A database tracks legal decisions that include AI hallucinations (1,825 cases so far).

“People expect AI systems to undergo rigorous testing to improve accuracy and reliability and to rid any technology deployed in the market of biases and discrimination,” said Williamson. “Unless AI developers take strong, continuing, affirmative steps to debias their AI systems, those systems will learn and reproduce inaccurate and harmful content.”

The NCLC comments also criticized the FTC’s suggestion that its policy proposal could preempt state laws protecting residents from the abuses of discriminatory and biased use of AI, calling it an “unsupported federal regulatory scheme.” NCLC signed on to coalition comments, led by the Lawyer’s Committee, in opposition to the proposed policy statement. 

Related Resources

• NCLC Comments on the FTC’s Proposed Policy Statement Concerning the Suppression of Accuracy in AI Systems, July 31, 2026
• Lawyers Committee Comments on FTC Proposed Policy Statement on AI, July 31, 2026
• Civil Rights Letter to the House Democratic AI Commission, July 16, 2026
• AI Order Seeks to Block States from Protecting People from Big Tech Risks, Dec. 12, 2025