Proposed Regulations Target Political Use of AI Deepfakes and Synthetic Media

As regulators worldwide grapple with the rapid proliferation of AI-generated deepfakes, policymakers face critical decisions about how to protect electoral processes without stifling free expression. The challenge has become increasingly urgent as sophisticated artificial intelligence makes it easier to create convincing fake videos, images, and audio that can mislead voters and potentially undermine democratic institutions.

Several fundamental questions should guide any regulatory approach, experts say. These include defining clear goals for regulation, determining which types of content should be covered, establishing appropriate enforcement mechanisms, and deciding who should be subject to these rules.

The most compelling rationale for regulation is the preservation of an informed electorate. As the Supreme Court noted in the landmark 1976 campaign finance case Buckley v. Valeo, “the ability of the citizenry to make informed choices among candidates for office is essential” to a functioning democracy. Deepfakes of candidates or events can directly threaten this fundamental interest.

Safeguarding the electoral process itself represents another powerful justification. Synthetic media that misleads voters about when, where, and how to cast their ballots—or that falsely depicts fraud to undermine confidence in election results—can have devastating consequences. The January 6, 2021 attack on the U.S. Capitol stands as a stark reminder of how electoral disinformation can fuel violence.

Protection of candidates and election workers provides a third rationale, though this must be balanced against the need for robust political discourse. While public figures necessarily accept greater scrutiny, the targeted harassment of election officials has reached alarming levels in recent years. AI-generated content could intensify these threats.

The scope of regulation presents another complex challenge. Some approaches target all “materially deceptive media” regardless of how they were created, while others focus specifically on synthetic content produced using artificial intelligence. A California law passed in 2019, for example, regulates manipulated candidate images distributed within 60 days of an election when intended to deceive voters.

Most experts recommend covering images, videos, and audio, as any of these formats can effectively mislead the public. Text-based synthetic content presents additional considerations, with some states like New York including AI-generated text in proposed regulations.

For regulatory approaches, policymakers have generally pursued two strategies: mandating disclosure or banning certain content outright. Disclosure requirements—such as disclaimers identifying AI-generated content—tend to face fewer constitutional challenges since they don’t prevent speech. Several states and proposed federal legislation like the AI Disclosure Act focus on this approach.

However, some jurisdictions have implemented targeted prohibitions. Texas and Minnesota have enacted bans on certain categories of deepfakes intended to influence elections, with these restrictions limited to specific pre-election windows (30 days in Texas, 90 days in Minnesota). Such prohibitions may be justified for content with little redeeming value, such as deliberate attempts to suppress voting.

The question of who should be regulated remains equally important. While most proposed regulations target candidates, political action committees, and others creating and distributing election-related deepfakes, some experts advocate for platform accountability as well. The bipartisan Honest Ads Act provides one potential model, requiring major online platforms to maintain public records of political ad purchases and ensure appropriate disclaimers.

As elections worldwide face this emerging threat, the challenge for policymakers is clear: develop targeted responses that mitigate harm without unduly restricting legitimate expression. While addressing synthetic media is crucial, experts caution that deepfakes ultimately represent a “threat amplifier” for existing challenges to electoral integrity rather than an entirely new problem.

With major elections approaching in dozens of countries in 2024, the urgency of thoughtful regulation has never been greater. Finding the right balance between protecting democratic processes and preserving free speech will be essential as societies navigate this rapidly evolving technological landscape.