By: Sarah Peck
As the 2026 election cycle approaches, lawmakers and regulators are moving quickly to address a growing threat: the use of artificial intelligence to create deceptive political content. In recent months, federal agencies and state governments have taken steps to regulate so-called “deepfakes”—AI-generated audio, images, and videos that can convincingly mimic real people.[1] These efforts reflect mounting concern that synthetic media could be used to mislead voters, spread disinformation, or undermine confidence in democratic processes.
The urgency is not theoretical. In 2024, voters in New Hampshire received robocalls featuring an AI-generated voice resembling President Joe Biden, urging them not to vote in the state’s primary.[2] The incident prompted investigations and enforcement actions, including criminal charges against a political consultant accused of orchestrating the scheme.[3] These developments signal that AI-generated political deception is no longer a hypothetical risk, but a present and evolving challenge.
Federal regulators have already begun to act. In February 2024, the Federal Communications Commission issued a declaratory ruling stating that AI-generated voices used in robocalls qualify as “artificial” under the Telephone Consumer Protection Act (TCPA), allowing such calls to be restricted or penalized under existing law.[4] The agency later proposed a multi-million dollar fine against those responsible for the New Hampshire robocalls, further demonstrating its willingness to enforce these rules.[5]
At the state level, legislative activity has accelerated. States such as California and Texas have enacted or expanded laws targeting deceptive synthetic media in political communications, including requirements for disclosure and restrictions on materially misleading content near elections.[6] In 2025 and 2026, additional states have introduced similar measures, creating a rapidly evolving patchwork of regulations aimed at curbing AI-generated election interference.[7]
The technology driving these developments has advanced rapidly. Voice cloning tools can now generate realistic speech from only a short audio sample, lowering the barrier to creating convincing impersonations.[8] As these tools become more accessible, the potential for misuse increases, allowing individuals with limited technical expertise to create synthetic media capable of reaching large audiences.
Despite the momentum toward regulation, significant constitutional challenges remain. Political speech lies at the core of First Amendment protection, and courts have consistently applied strict scrutiny to laws that restrict it.[9] Efforts to regulate AI-generated political content must therefore be carefully tailored to avoid infringing on protected expression, including satire, parody, and legitimate political messaging.
This creates a difficult legal question: should regulation focus on the use of AI itself, or on the deceptive effects of its outputs? Broad restrictions on AI-generated political content risk sweeping in lawful speech, while narrower rules targeting deception may be difficult to enforce in fast-moving digital environments. The speed and scale of AI-generated content make it particularly challenging to detect and address harmful uses before they reach voters.
The current wave of enforcement and legislation suggests that regulators are attempting to strike a middle ground—targeting clearly deceptive uses of AI-generated content while preserving room for lawful expression. However, as deepfake technology continues to evolve, these efforts may struggle to keep pace. The speed at which synthetic media can be created and disseminated presents a fundamental challenge to traditional legal approaches, which often rely on after-the-fact enforcement.
Ultimately, the question is not whether AI-generated deepfakes will play a role in future elections—they almost certainly will. The more pressing issue is whether existing legal frameworks can effectively address the risks they pose. As courts begin to confront these questions, their decisions will shape not only the regulation of AI, but also the future of political communication in an era where seeing—and hearing—is no longer believing.
Sources
- Reuters, U.S. states ramp up regulation of AI deepfakes ahead of 2026 elections (Feb. 14, 2025), https://www.reuters.com/technology/artificial-intelligence/us-states-ramp-up-ai-deepfake-regulation-ahead-2026-elections-2025-02-14/ [hereinafter Deepfakes].
- Reuters, Fake ‘Biden’ robocall tells New Hampshire Democrats to stay home (Jan. 22, 2024), https://www.reuters.com/world/us/fake-biden-robo-call-tells-new-hampshire-voters-stay-home-2024-01-22/.
- Press Release, N.H. Dep’t of Justice, Steven Kramer Charged with Voter Suppression Over AI-Generated President Biden Robocalls (May 23, 2024), https://www.doj.nh.gov/news-and-media/steven-kramer-charged-voter-suppression-over-ai-generated-president-biden-robocalls.
- In re Implications of Artificial Intelligence Technologies on Protecting Consumers from Unwanted Robocalls and Robotexts, Declaratory Ruling, FCC 24-17 (Feb. 8, 2024), https://docs.fcc.gov/public/attachments/FCC-24-17A1.pdf.
- In the Matter of Steve Kramer, Notice of Apparent Liability for Forfeiture, FCC 24-59 (adopted May 23, 2024; released May 24, 2024), https://docs.fcc.gov/public/attachments/FCC-24-59A1.pdf.
- Cal. Elec. Code § 20010 (West 2024); Tex. Elec. Code Ann. § 255.004 (West 2023).
- Deepfakes¸ supra note 1.
- FTC Office of Technology, Preventing the Harms of AI-Enabled Voice Cloning (Nov. 16, 2023), https://www.ftc.gov/policy/advocacy-research/tech-at-ftc/2023/11/preventing-harms-ai-enabled-voice-cloning.
9. Brown v. Ent. Merchants Ass’n, 564 U.S. 786 (2011).