Loading issue research
Require consent and compensation for a real person's likeness or voice, and disclosure of synthetic commercial content, through state law and industry standards, since federal enforcement here reversed in 2025.
Verification Status
AI-researched, unverifiedLast Reviewed
Jul 4, 2026
Cited Sources
7
What is failing, what we would change, and the conclusion we are willing to defend.
This issue is distinct from AI-10, which covers whether a model may train on copyrighted creative work at all. This issue covers what happens on the output side: should AI-generated commercial content be disclosed, and does using a real, identifiable person's voice or likeness require their consent and compensation. The federal regulatory picture here has moved backward: in December 2025, the FTC vacated its own model AI-fake-review enforcement action, reasoning that restricting a technology over potential misuse "unduly burdens AI innovation." Progress on both disclosure and likeness-consent has instead come from states and negotiated industry agreements.
Support state-level AI-content disclosure laws, following New York's requirement to disclose AI-generated "synthetic performers" in paid advertising, and California's requirement that large providers offer free detection tools. Federal enforcement has already reversed its own position in this area.
Support a federal floor requiring consent and compensation for commercial use of a real person's digital voice or likeness. This isn't a hypothetical ask: the NO FAKES Act already does exactly this, and its sponsor list crosses the aisle — Rep. Salazar (R) and Rep. Dean (D) in the House, Sen. Blackburn (R) and Sen. Coons (D) in the Senate. Model the federal floor on the entertainment-industry union's negotiated consent-and-compensation framework and New York's estate-consent requirement for deceased performers. This protects performance and likeness, a distinct right from AI-10's training-data compensation for training-corpus contribution.
Support continued industry adoption of content-provenance standards as the primary practical disclosure mechanism, while being honest that the standard breaks down once content crosses a platform that strips metadata during re-encoding, the same limitation already named in AI-13.
Don't restrict or cap AI-generated creative content itself. Market disruption in several creative professions is documented below, but it doesn't currently justify prohibition any more than AI-11's labor data justifies a broader ban on AI-assisted work. Support targeted disclosure and consent requirements instead.
Support platform-level accountability for undisclosed synthetic content that could deceive consumers about a work's authorship or a performer's participation, a market-integrity concern narrower than and distinct from the election-specific deception already covered in AI-13.
Require platforms that host AI-generated content at commercial scale — marketplaces, video platforms, app stores — to implement working provenance and quality enforcement with measurable controls: the same identity-verification standard a major bookseller now applies to publishing accounts should be the floor other large platforms are held to, and platforms that knowingly profit from mass low-quality synthetic content impersonating authorship (fake books mimicking an author's catalog, AI-generated local "news" sites impersonating journalism) should face accountability for it before press coverage forces the issue. Scale the obligation to platform size — a revenue or user-count threshold, consistent with the safe-harbor approach this platform already uses in AI-05 and AI-10 — so the standard targets the marketplaces large enough to be a meaningful vector for this harm, not every small platform regardless of exposure.
Require consent and compensation for a real person's likeness or voice, and disclosure of synthetic commercial content, through state law and industry standards, since federal enforcement here reversed in 2025.
Turn frustration into useful pressure.
If this position misses evidence or a lived consequence, challenge it. If it holds up, help test it locally and connect it to the issues around it.