A trio of US House lawmakers introduced legislation that would mandate that AI-generated images, video and audio contain machine-readable disclosures identifying the source of the content.
The Protecting Consumers from Deceptive AI Act — reintroduced April 24 by Reps. Valerie Foushee (D-NC), Don Beyer (D-Va.) and James Moylan (R-Guam) — requires the National Institute of Standards and Technology (NIST) to establish task forces within 90 days of enactment to develop technical standards covering watermarking, digital fingerprinting and content provenance metadata for AI-generated content. The Federal Trade Commission would be empowered to bring enforcement actions against companies that fall short.
The bill would put the sharpest obligations on AI application providers and the large platforms that host their outputs. Covered online platforms, which the proposal defines as services generating at least $50 million in annual revenue or drawing 25 million or more monthly active users in at least three of the prior 12 months, would be barred from stripping AI-origin disclosures from content and required to surface that information to users.
"Deepfakes and AI-generated audio and visual content [pose] major risks to consumers, our elections, and public trust. Clear labeling and transparency of this content must be required so Americans can distinguish what images, audio, and videos are artificially generated," Foushee said. "As the spread of deceptive AI content fuels misinformation and raises serious civil rights concerns, particularly for those that are disproportionally targeted online, this bill is an important step towards protecting creators and their work and ensuring generative AI systems are not used to hurt our communities."
Violations would be treated as unfair or deceptive acts under the FTC Act, giving the commission civil penalty authority. The FTC would have two years after enactment to promulgate implementing rules. The disclosure requirements for AI application providers and platforms would not become effective until 90 days after those rules take effect, meaning real-world obligations are at least 2.5 years out at the earliest.
The bill would direct NIST to develop guidelines for identifying and labeling AI-written text, but it does not seek to impose the same hard disclosure mandate to audio and video content.
The legislation also includes a self-regulatory safe harbor. Companies that comply with FTC-approved industry guidelines would be deemed compliant with the proposed law's requirements. The commission would have 180 days to act on any safe harbor request.
NIST task forces would be required to submit technical recommendations to the institute's director within 270 days of establishment and file annual reports to the House Science, Space and Technology Committee, the House Energy and Commerce Committee, and the Senate Commerce, Science and Transportation Committee for five years.
The bill drew endorsements from Adobe Inc., the Authors Guild, Encode AI, IEEE-USA, the Society of Composers and Lyricists and the American Society for Collective Rights Licensing.
"Generative AI can be used to create highly realistic content that is difficult or impossible to distinguish from authentic material, including deepfakes," said Cynthia Rudin, a professor of computer science at Duke University. "We are not interested in being deceived, and we are not interested in our children being deceived either. This is a matter of personal safety, national security, and maintaining a properly informed electorate."
A spokesperson for Foushee did not immediately respond to a request for comment.
What's happening this week?
Below is a list of hearings, webinars and other tech, media and telecom-related events taking place virtually and in person in the nation's capital and beyond this week:
April 28
➤ Senate Appropriations Subcommittee on Commerce, Justice, Science, and Related Agencies: Hearings to examine proposed budget estimates and justification for fiscal year 2027 for the National Aeronautics and Space Administration.
April 29
➤ House Homeland Security Subcommittee on Cybersecurity and Infrastructure Protection: Data Centers, Telecommunications Networks, and Space-Based Systems — Modernizing DHS's SRMA Role for the Communications and IT Sectors
➤ House Energy and Commerce Subcommittee on Energy: AI and the Grid — Meeting Growing Power Demand While Protecting Ratepayers
April 30
➤ Information Technology and Innovation Foundation: Strengthening America's Edge in Priority Technologies of the 21st Century
➤ Stanford Freeman Spogli Institute for International Studies: Visions of Taiwan's Future — Navigating Paths Through Democracy, Technology, and Culture
➤ Federal Communications Commission: Open Meeting
