FTC PUTS AI COMPANIES ON NOTICE — SECRETLY STEERING YOUR MODEL’S OUTPUTS TO HIT UNDISCLOSED GOALS IS NOW ILLEGAL UNDER FEDERAL LAW The Federal Trade Commission released a proposed policy statement this month targeting what it calls the suppression of accuracy in AI systems, putting every major AI company on notice that secretly manipulating model outputs to serve undisclosed ideological or commercial objectives could constitute a deceptive trade practice under Section 5 of the FTC Act. Public comment closes July 31. The FTC’s core argument is straightforward and damaging. AI companies have explicitly and implicitly told consumers their systems are designed to produce the best, most accurate, most faithful output possible. Consumers rely on that promise. They accept AI outputs without independent fact-checking more than 90 percent of the time, according to the agency. If a company is covertly tuning its model to avoid certain conclusions, favor certain viewpoints, or serve corporate interests it has never disclosed, it is lying to the people paying for it. The proposed statement stops short of naming specific companies or models, but the implications sweep across the industry. Every AI system that generates recommendations, answers questions, or writes content for consumers is potentially covered. The agency is specifically concerned with cases where models are tuned to avoid disparate impact liability under state law, which it argues can produce its own form of distorted output. Comments are open through July 31. If finalized, this statement gives the FTC a clear legal hook to pursue enforcement against AI companies whose outputs diverge systematically from their stated accuracy commitments. Keywords: FTC AI policy, AI accuracy regulation, FTC Section 5 AI, AI output manipulation regulation