What is the NO FAKES Act, and what does it actually do?
The NO FAKES Act — formally the Nurture Originals, Foster Art, and Keep Entertainment Safe Act of 2026 — is a bipartisan federal bill that gives every individual a property right over AI-generated “digital replicas” of their voice or visual likeness, letting them license, condition, or refuse consent the same way they would license footage or a trademark.
- A federal property right — Every individual gets a right in their own voice and visual likeness that they can license, meaning a creator or their agency can grant, condition, or withhold consent for an AI replica.
- Liability for creation and distribution — Both the party that creates an unauthorized digital replica and the party that distributes it can be held liable, exposing the brand that commissions the replica and the platform running the ad, not just the model builder.
- A DMCA-style notice-and-takedown process — A person whose likeness is used without consent can demand removal directly from a platform without filing a lawsuit first. Whoever posted the content can file a counter-notice; if no lawsuit follows within roughly 14 days, the content is restored.
- Tiered statutory damages — Current bill text puts statutory damages at roughly $5,000 per unauthorized replica for individual violators, up to $25,000 per work for platforms that make good-faith takedown efforts, and as high as $750,000 per work for platforms that don't.
- Carve-outs — First Amendment uses (news, parody, criticism, commentary) are excluded, along with nonprofit libraries, archives, and educational institutions doing non-commercial research, plus technical clarifications for streaming music platforms.
The damages figures above come from bill-tracking summaries of the current text, not an enacted statute — no statute exists yet, and the exact numbers could shift as the bill moves through the full Senate and any House reconciliation.
How did the NO FAKES Act get from a 2023 draft to a 2026 committee vote?
The NO FAKES Act has now been through four legislative stages across three years without a floor vote, before its first unanimous committee advancement on June 18, 2026 — a pattern that signals bipartisan consensus on the core idea, even though the bill is not yet law.
| Date | Stage | Detail |
|---|---|---|
| 2023 | Discussion draft | Bipartisan draft from Sens. Blackburn, Coons, Klobuchar, and Tillis. |
| July 2024 | First Senate introduction | Formal bill introduced; a House companion followed in September 2024. |
| April 2025 | Reintroduced | Reintroduced in both chambers after the 2024 session ended without a floor vote. |
| May 20, 2026 | Revised reintroduction | Blackburn, Coons, Salazar, and Dean introduced a strengthened version, S. 4591. |
| June 18, 2026 | Committee vote | Senate Judiciary Committee advanced S. 4591 by unanimous voice vote. |
Per Senator Blackburn’s June 2026 office statement, the bill now goes to the full Senate for a floor vote. The House companion has not yet advanced through committee, and no effective date is specified in the current text.
Who is liable under the NO FAKES Act, and what are the penalties?
Both the creator and the distributor of an unauthorized digital replica can be liable, with damages scaled by role: individual violators face a flat per-replica figure, while platforms face a tiered figure that depends on whether they made a good-faith takedown effort.
| Party | Statutory damages | Condition |
|---|---|---|
| Individual violator | ~$5,000 per unauthorized replica | Or actual damages, whichever is greater. |
| Good-faith platform | Up to ~$25,000 per work | Made a documented, timely takedown effort. |
| Non-compliant platform | Capped around ~$750,000 per work | Did not make a good-faith takedown effort. |
These figures come from bill-tracking summaries of the current text (Congress.gov, S. 4591), not an enacted statute — treat them as directional until a final version is signed into law. The notice-and- takedown mechanism itself moves in roughly 14 days, per Holland & Knight’s June 2026 analysis, which is days, not the months a right-of-publicity lawsuit would take today.
What’s the difference between AI-assisted content and an AI digital replica?
An AI digital replica represents a specific, identifiable real person’s voice or face; ordinary AI-assisted creative — image touch-ups, AI-written scripts, fully synthetic personas with no real-world counterpart — does not, and sits outside the NO FAKES Act’s scope entirely.
That distinction matters operationally because teams already running AI-generated ad variations or comparing AI UGC against creator UGC need to tag each asset by which bucket it falls into. A face swap or voice clone trained on a real creator’s past content needs a documented consent record before it ships; a fully synthetic AI spokesperson does not, because there is no real person whose likeness right applies.
What should a creator marketing team do before the bill is enacted?
Five concrete steps close most of the gap: inventory existing AI assets that replicate a real creator, confirm documented consent for each one, update the standard contract template, assign a record owner, and build a same-day takedown response — none of which require the bill to be law first.
- Inventory every AI-generated asset in production — Flag which ones replicate a real creator's voice or face versus which are fully synthetic, so the two buckets get different approval paths.
- Confirm documented consent on every flagged asset — Locate the specific grant covering that specific AI use, not a general 'content featuring the creator' license that predates the AI-generation step.
- Update the standard creator-agreement template — Spell out which digital-replica type is permitted (voice only, face only, full likeness), what it can be used for, for how long, and whether consent is revocable.
- Assign an owner for likeness-consent records — The same way someone already owns usage-rights tracking today, so consent has an accountable owner rather than living in a legal team's inbox.
- Build a same-day takedown response — A notice-and-takedown window running in days, not months, needs a documented path to pull an asset immediately, not a plan to route it through legal review first.
How does likeness consent fit into a workflow that already tracks usage rights?
For a team whose AI-generated creator assets already carry a structured usage-rights record, likeness consent is one more field on that same record, not a new compliance system built from scratch.
Storika’s campaign evidence object already attaches a source creator, an approved use, and an expiration window to every AI-generated ad variation that moves through a campaign — the same three-field shape that usage-rights tracking relies on. Likeness consent for a digital replica is a fourth field with that identical shape: who consented, to what specific AI use, until when. Teams that have already built that record for usage rights add likeness consent as an incremental field; teams still tracking AI assets in spreadsheets are building both systems at once.
The same artifact-provenance record that already tracks which prompt, model, and source asset produced a given piece of AI creative is the natural place to also record whose likeness it uses and under what consent — provenance and likeness consent are the same kind of fact about the same asset, tracked at the same point in the workflow.
Frequently asked questions
Is the NO FAKES Act law yet?
No. As of this writing, it has cleared the Senate Judiciary Committee on a unanimous voice vote (June 18, 2026) and moves to the full Senate. It still needs to pass the Senate, pass or reconcile with a House version, and be signed into law before it takes effect.
Does the NO FAKES Act apply to AI-generated content that doesn't look like a real person?
No. The bill targets digital replicas of a specific individual's voice or visual likeness. Fully synthetic AI personas with no real-world counterpart fall outside its scope, regardless of how the underlying content was produced.
Who can be held liable under the NO FAKES Act?
Based on current bill text, both the party that creates an unauthorized digital replica and the party that distributes it can face liability, which reaches brands, agencies, and the platforms that run the content, not only whoever built the underlying AI model.
What should creator marketing teams do right now, before the bill is enacted?
Update creator agreements to name AI-replica use explicitly, start tracking likeness consent the same way usage rights are tracked today, and separate 'AI-assisted content' from 'AI replica of a specific real creator' internally so the right approval checks apply to each.
Does the NO FAKES Act replace state right-of-publicity laws?
No. The bill creates a federal standard specifically for AI-generated digital replicas. It does not eliminate existing state right-of-publicity protections, which continue to apply alongside it once the federal bill takes effect.
How is this different from ordinary AI-generated UGC or AI-assisted ad creative?
Ordinary AI-assisted creative, such as image touch-ups or fully synthetic AI personas, doesn't replicate a specific real person, so it sits outside the bill's target. The trigger is whether the output represents a named, identifiable creator's actual voice or face without that creator's consent.
Related reading
This is the first Storika guide covering AI likeness and digital-replica law specifically — pair it with kidfluencer compliance (child-specific creator law), the general FTC disclosure workflow, AI-generated ad variations, AI UGC vs. creator UGC, usage-rights tracking, and AI outreach artifact provenance so likeness consent, usage rights, disclosure, and provenance all run as one record instead of four disconnected checks.
Sources
- Blackburn, Coons Bipartisan Bill to Protect Individuals and Creators From Deepfakes Passes Senate Judiciary Committee — Senator Marsha Blackburn, official press release (June 18, 2026)
- Rep. Salazar’s NO FAKES Act Advances Out of Senate Judiciary Committee With Unanimous Support — Representative Maria Salazar, official press release (June 18, 2026)
- Blackburn, Coons, Salazar, Dean, Colleagues Introduce Revised Version of NO FAKES Act — Senator Marsha Blackburn, official press release (May 20, 2026)
- S.4591 — NO FAKES Act of 2026 — Congress.gov, Library of Congress (bill text and status)
- Senate Committee Advances Bill to Protect Name, Image, Likeness and Voice Against Unauthorized AI Use — Holland & Knight, legal analysis (June 2026)
- Navigating AI Liability Under the Revised NO FAKES Act — Fenwick, legal analysis