Congress Weighs New Deepfake Legislation as AI-Generated Content Floods 2026 Campaigns

Congressional hearings have begun on new legislation targeting AI-generated deepfakes in political campaigns, with bipartisan support emerging for mandatory disclosure requirements. The hearings come as AI-generated robocalls, synthetic video, and fabricated endorsements proliferate ahead of the 2026 midterm elections.

The Deepfake Problem

The 2026 election cycle has seen a dramatic increase in AI-generated political content: synthetic video clips showing candidates making statements they never made, AI-generated voice calls impersonating candidates and election officials, and fabricated images designed to create false impressions of candidate behavior. The content is increasingly difficult to distinguish from authentic media, and its distribution through social media platforms means it can reach millions of voters before fact-checkers can respond.

Proposed Legislation

The proposed bill would require mandatory disclosure when AI is used to generate or substantially modify political advertising and campaign communications, create a federal standard for AI content watermarking that platforms would be required to detect and label, establish criminal penalties for the knowing distribution of undisclosed AI-generated content intended to influence elections, and direct the FEC to issue guidance on AI use in political campaigns. Notably, the bill has bipartisan co-sponsors, reflecting rare agreement that AI-generated election interference is a nonpartisan threat.

Industry Response

Technology companies have expressed support for disclosure requirements while opposing more restrictive provisions that could affect legitimate AI uses in content creation. The watermarking requirement is technically challenging: current AI watermarking systems can be removed or circumvented by determined actors, and no watermarking standard has been widely adopted across AI model providers. The legislation's effectiveness will depend on whether watermarking technology can be made robust enough to survive adversarial removal attempts.