Protect Elections from Deceptive AI Act
- Bill Number
- H.R. 5272
- Origin Chamber
- House
- Congress
- 119th Congress, Session 1
- Policy Area
- Government Operations and Politics
- Status
- Introduced
- Latest Action
- 2025-09-10: Referred to the House Committee on House Administration.
- Last Updated
- 2026-06-09T08:05:43Z
AI-Generated Summary
Purpose of the Legislation
The "Protect Elections from Deceptive AI Act" (H.R. 5272) aims to safeguard federal elections by banning the spread of fake audio or video content created by artificial intelligence (AI) that could mislead voters about candidates. It targets "deepfakes" – AI-generated media that looks real but falsely depicts a candidate's appearance, speech, or actions – to prevent interference in elections or fundraising.
Key Provisions
- Definitions:
- Covered individual: Any candidate running for federal office (e.g., President, Congress).
- Deceptive AI-generated audio or visual media: AI-produced images, audio, or videos (using techniques like machine learning) that appear authentic but would mislead a reasonable person into believing they show real events, speech, or conduct of a candidate that didn't actually happen. This is judged based on the media's quality and how/where it's shared.
- Federal election activity: Refers to efforts tied to influencing federal elections, as defined in existing election law.
- Prohibition:
- Individuals, political committees (groups that raise/spend money on elections), or other entities cannot knowingly distribute this deceptive AI media about a covered individual if the intent is to influence a federal election or raise funds.
- The ban applies specifically during federal election activities.
- Exceptions (where the prohibition does not apply):
- Radio/TV stations, cable/satellite providers, or streaming services, if they air the content as part of legitimate news (e.g., newscasts, interviews, documentaries) and clearly disclose (in a way that's easy to hear or see) that its authenticity is questionable.
- Newspapers, magazines, or online publications that regularly cover news, if they publish the content but explicitly state it doesn't accurately show the candidate's real speech or actions.
- Content that's clearly satire or parody (humorous imitations not meant to deceive).
- Enforcement and Remedies:
- Civil lawsuits: Affected candidates can sue for court orders to stop distribution (injunctive relief) or other equitable remedies. These cases get priority handling in court.
- Damages: Candidates can seek financial compensation (general or special damages) from violators. Courts may also award attorney's fees and costs to the winner. This doesn't limit other legal options.
- Burden of proof: Plaintiffs (the candidates suing) must prove the violation with "clear and convincing evidence" – a higher standard than usual, meaning strong, persuasive proof.
- Any violation of this law counts as "defamation per se" (automatic libel/slander without needing to prove harm to reputation) for defamation lawsuits.
- Severability: If any part of the law is ruled invalid by a court, the rest remains in effect.
Significant Changes to Existing Law
- Amends the Federal Election Campaign Act of 1971 (a key U.S. law regulating campaign finance and activities) by adding a new Section 325 specifically targeting AI-generated deceptive media.
- Introduces the first federal prohibition on distributing deepfakes in elections, with tailored civil remedies (e.g., priority court handling, defamation per se status) not previously available for this type of content.
- No criminal penalties; focuses on civil actions rather than expanding government enforcement like fines from the Federal Election Commission.
Potential Impacts
- On Citizens/Voters: Reduces the risk of election misinformation from AI deepfakes, helping voters make informed choices without being tricked by fake candidate content. Could increase trust in elections but might limit some online sharing of AI media.
- On Government Agencies: Minimal direct impact, as enforcement relies on private lawsuits rather than agency action (e.g., no new role for the Federal Election Commission). Courts will handle cases, potentially increasing their workload.
- On International Relations: None directly addressed; the law focuses on U.S. federal elections and doesn't mention foreign actors, though it could indirectly deter foreign interference via AI if content targets U.S. candidates.
- Broader Effects: May encourage AI developers and platforms to add watermarks or detection tools for election-related content, promoting transparency in digital media.
Main Stakeholders Affected
- Candidates for Federal Office: Primary beneficiaries, as they can sue to stop harmful deepfakes and seek damages; also protected from misuse of their likeness.
- Political Committees and Entities: Restricted from using or sharing deceptive AI media in campaigns or fundraising; could face lawsuits if they violate the rules.
- Media Outlets (TV, Radio, Print, Online): Can still cover AI content under exceptions if they disclose issues, but must be cautious to avoid liability.
- AI Creators, Tech Companies, and Distributors: Limited in creating/sharing election-related deepfakes; may need to implement safeguards like authenticity labels.
- Voters and the Public: Indirectly protected from deception, but exceptions ensure access to news and satire.
Notable Legal, Constitutional, or Political Implications
- Legal: Establishes a clear civil framework for addressing AI in elections, with a high proof burden to avoid frivolous suits. The defamation per se rule strengthens candidates' recourse but could lead to more litigation over what counts as "deceptive."
- Constitutional: Balances First Amendment free speech rights by exempting news, commentary, and parody, reducing risks of overbroad censorship. However, the prohibition on intent-based distribution might face challenges if seen as vague or chilling political expression.
- Political: Promotes election integrity amid rising AI concerns, potentially influencing campaign strategies by discouraging deepfake tactics. Bipartisan sponsorship (Democrats and Republicans) suggests broad support, but implementation could spark debates on tech regulation vs. innovation.
This summary was generated by AI and may contain inaccuracies. Refer to the official source document for the authoritative text.
Sponsor
Cosponsors (5)
Rep. Fitzpatrick, Brian K. [R-PA-1], Rep. Houlahan, Chrissy [D-PA-6], Rep. Gonzales, Tony [R-TX-23], Rep. Alford, Mark [R-MO-4], Rep. Hoyle, Val T. [D-OR-4]
Recent Actions
- 2025-09-10: Referred to the House Committee on House Administration.
- 2025-09-10: Introduced in House
- 2025-09-10: Introduced in House
Bill Versions
- Protect Elections from Deceptive AI Act — issued 2025-09-10 — PDF (7 pages)