New law makes it easier to sue over deepfake endorsements—but shields platforms

What it does

This bill amends federal trademark law to make it easier for living individuals to sue when AI-generated digital replicas of their face, voice, or other identifying characteristics are used in commerce without permission—particularly in fake endorsements and fraudulent advertisements. It creates a legal presumption that such unauthorized use causes consumer confusion, shifting the burden to defendants to prove otherwise, while carving out protections for expressive works (films, music) and First Amendment speech.

Why we flagged it

The bill's operative mechanism is a rebuttable presumption of consumer confusion in trademark cases involving unauthorized AI-generated digital replicas used in commerce. This is a targeted consumer-protection measure addressing a specific, documented harm (deepfake endorsements and fraud), not a broad deregulation or industry carve-out.

What the text implies

• The rebuttable presumption may shift litigation costs: defendants must now affirmatively disprove confusion rather than plaintiffs proving it, potentially increasing settlement pressure on platforms and AI companies even for borderline cases.
• The exclusion of 'indirect liability' claims (subsection 5(A)) may shield platforms and AI service providers from liability if they host or distribute deepfakes, even if they knowingly facilitate the abuse—the presumption applies only to direct users.
• The 'relevant industry or market' definition ties the presumption to the goods/services being promoted, not the person's actual fame, meaning a non-celebrity can invoke the presumption if their digital replica is used to sell a product in a niche market.
• Section 230(e) carve-out (subsection 7) may create a new exception to Section 230 immunity for platforms, but only if a plaintiff invokes the rebuttable presumption—this could incentivize plaintiffs to use this mechanism strategically to bypass platform immunity.
• The 'expressive work' carve-out (subsection 5(B)) is undefined and may generate litigation over whether a deepfake in a film, music video, or satirical advertisement qualifies as protected expression.

Section numbers refer to the bill text the analysis read — linked under Primary records below.

Who it affects

The bill protects ordinary citizens from a concrete, growing harm: fraudulent AI-generated endorsements that damage reputation and deceive consumers into buying dangerous products. By lowering the evidentiary burden for non-famous individuals to prove consumer confusion, it democratizes access to trademark remedies. The carve-outs for expressive works and First Amendment speech preserve legitimate creative and political speech.

Who stands to gain

• Individuals (living persons) harmed by deepfake endorsements and fraud
• Plaintiff's bar (trademark and IP litigation attorneys)
• Potentially: AI detection and authentication technology vendors

Named in the bill

Trademark Act of 1946, Section 43(a) of Trademark Act, Section 230 of Communications Act of 1934, Section 32 and 43(c) of Trademark Act, Section 501 of Title 17 (Copyright), Section 271 of Title 35 (Patents), First Amendment, House Committee on the Judiciary

Where it stands

• Aug 3, 2026 — Introduced · Congress.gov: “Introduced in House”
• Aug 3, 2026 — Referred to House Committee on the Judiciary · Congress.gov: “Referred to the House Committee on the Judiciary”

Dates and quoted wording are Congress.gov's action record; the timeline shows status changes, not every procedural step.