300-plus comments on an FTC AI speech draft turn content policy into a compliance surface

A little-noticed Federal Trade Commission draft policy aimed at AI-generated speech that is, in the draft's words, distorted by undisclosed ideological objectives and by so-called equity has collected more than 300 public comments, and the filings show deep concern about the government's role in policing what models say ^[1][2]. The consequence for anyone shipping a model is that refusal behaviour and tone, long treated as internal trust-and-safety judgement, are now being argued over as an object of federal jurisdiction.

According to Fast Company's reading of the docket, many influential industry, legal and advocacy groups argued that the FTC is setting up its own lever to control what AI services say, and the draft was broadly panned by groups across the ideological spectrum ^[3]. The Center for Media and Digital Governance at the anti-monopoly Open Markets Institute argued that content controls can amount to an exploitation of power concentration, and compared the proposal to China's censorship of its own AI models, which it said systematically refuse to engage with Tiananmen Square, Taiwan, and criticism of the Communist Party ^[5]. That is an unusual pairing of opponents: the same instrument read as a speech risk by an antitrust group and, per the same reporting, defended by Innovation Council Action, a group linked to Trump administration technology advisors, on the grounds that it is necessary to preempt state-level AI policies and keep pace with China ^[7]. At least one filer therefore wants the rule for reasons that have nothing to do with ideological distortion and everything to do with a single federal ceiling replacing state floors ^[13].

China is the load-bearing argument in both directions. A striking minority of the comments were preoccupied with it, and the reporting counts four distinct roles it plays in the record: threat, justification for light-touch regulation on competitiveness grounds, foil for American values, and accusation that the administration is itself pursuing censorship ^[4][14]. The Foundation for Defense of Democracies took the inverse position from Open Markets, arguing that content oversight and transparency are needed to counter bias inherent to Chinese models that are gaining use inside the United States ^[6]. The FTC had already staked out that ground in the draft, noting that geopolitical rivals are investing heavily in the sphere and hope to inject their own companies and values into the marketplace ^[8]. The article's author argues that fear of the world's leading practitioner of content control could end up giving staying power to a policy that critics see as a First Amendment threat ^[9].

For operators, the transferable point is the one the author lands on: policymakers are asking private companies and individuals to steer speech in an area where AI services have clear technical capability, even if equating the FTC's draft with China's broad content controls would be a stretch ^[12]. Compare the direction of travel: China's 2023 measures direct AI services to protect the nation's image and uphold socialist core values, and its 2025 AI Safety Governance Framework 2.0 highlighted information content risks ^[10].

Watch whether the preemption rationale, not the ideology rationale, is what survives into any final text ^[7], and whether the competitiveness framing the FTC itself introduced ^[8] becomes the durable justification ^[9].