The U.S. Senate Judiciary Committee has unanimously advanced legislation that would give every American the right to control how their voice and face are used in AI-generated deepfakes. The NO FAKES Act, formally titled the Nurture Originals, Foster Art, and Keep Entertainment Safe Act of 2026, passed on June 18 and now moves to the full Senate for a vote.

This marks the first major federal effort to address a growing problem: generative AI systems that can convincingly mimic anyone's voice or appearance without permission. From fraudulent celebrity endorsements to deepfake videos of public figures, these AI-generated replicas have already caused real harm. The bill aims to create a uniform national standard instead of forcing people to navigate a patchwork of state laws.

What Would the NO FAKES Act Actually Do?

At its core, the legislation creates a new federal intellectual property right that belongs to you. Whether you are a celebrity or a private citizen, you would own the right to control how your voice and visual likeness appear in digital replicas. This is a significant shift because it treats your voice and face as property you can license, protect, or keep private.

The bill establishes clear consequences for violations. Individuals and companies that produce or distribute unauthorized digital replicas could face liability. Online platforms that host such content and know it was not authorized could also be held responsible. This creates incentives for platforms to act quickly when someone reports a deepfake.

How Would People Remove Unauthorized Deepfakes?

The legislation borrows a proven framework from copyright law. It creates a notice-and-takedown process similar to the Digital Millennium Copyright Act (DMCA), which has governed online copyright disputes for decades. If you discover your voice or likeness has been used without permission, you can demand that the content be removed from platforms without having to file an expensive lawsuit first.

The process includes built-in protections against abuse. If someone files a counter-notice claiming the takedown was a mistake or that the replica was actually authorized, the platform must notify you. You then have 14 days to file a lawsuit, or the content gets restored. Anyone who knowingly files a false counter-notice faces penalties of $25,000 per violation or actual damages plus legal fees, whichever is greater.

What Exceptions Does the Bill Include?

The NO FAKES Act does not ban all AI-generated replicas. It includes important carve-outs for First Amendment protections and legitimate uses. The bill specifically excludes news reporting, parody, criticism, and similar forms of protected speech. It also creates exemptions for nonprofit libraries, archives, and accredited nonprofit educational institutions conducting non-commercial research.

These exceptions reflect a careful balance. Lawmakers recognized that some uses of digital replicas serve the public interest or artistic expression. A news organization reporting on a political figure, for example, could use a deepfake to illustrate a story without violating the law. A comedian could parody a public figure's voice for satire. The bill aims to protect individuals while preserving free speech.

How Does This Fit Into Broader AI Governance Trends?

The NO FAKES Act represents a shift in how U.S. policymakers approach AI regulation. Rather than waiting for a comprehensive federal AI law like the European Union's AI Act, Congress is addressing specific harms through targeted legislation. This approach allows lawmakers to move quickly on urgent problems while the broader regulatory landscape remains fragmented.

The United States currently lacks a single federal AI law comparable to Europe's comprehensive framework. Instead, the regulatory environment consists of executive orders, agency guidance, sector-specific rules, and state statutes. This patchwork means compliance obligations depend on where a company operates and what decisions the AI supports.

The NO FAKES Act also addresses a gap in existing state laws. While Tennessee has passed the ELVIS Act (Ensuring Likeness Voice and Image Security Act), creating state-level protections, most states have no uniform rules. The federal bill would preempt future state laws on this issue but would not eliminate existing state protections like Tennessee's law.

Steps for Compliance and Implementation

• Platform Responsibility: Online platforms must establish systems to receive, process, and respond to takedown notices within a reasonable timeframe, similar to existing DMCA procedures they already use for copyright claims.
• Creator Awareness: Content creators and AI developers need to understand that using someone's voice or likeness without consent could trigger liability, even if the deepfake is technically sophisticated or commercially valuable.
• Documentation and Consent: Companies developing AI systems that generate realistic replicas should maintain clear records of consent from individuals whose voices or likenesses are used in training data or final products.
• Monitoring and Enforcement: Rights holders will need to actively monitor platforms and the internet for unauthorized replicas, then file takedown notices when violations occur, similar to how copyright holders police digital content today.

What Happens Next?

The bill's unanimous passage out of committee is a strong signal of bipartisan support, but some senators expressed concerns about whether the legislation adequately protects legitimate First Amendment speech. The sponsors have agreed to work with these senators as the bill moves to the full Senate floor.

A companion bill has also been introduced in the House of Representatives, though the House Judiciary Committee has not yet taken action on it. If both chambers pass the legislation, it would become the first major federal law specifically addressing AI-generated deepfakes of people's voices and faces.

The timing reflects growing urgency around deepfakes. As generative AI tools become cheaper and easier to use, the risk of unauthorized replicas increases. The legislation aims to establish clear legal protections before deepfake technology becomes even more widespread and harder to detect.