With generative AI now widely accessible and increasingly used to manipulate electoral narratives, lawmakers face critical challenges in crafting election laws that punish misleading conduct narrowly rather than banning the underlying technology outright.
- AI-enabled misinformation in elections is a growing threat evidenced by documented cases.
- Broad bans on AI technology face constitutional free speech challenges and practical limits.
- Narrow, conduct-focused laws targeting specific deceptive acts show more legal promise.
What happened
The 2026 US midterm elections are the first national elections where generative AI tools—capable of producing convincing fake audio and video—are cheaply and widely accessible. Election officials have already noted cases such as manipulated ballot counting footage in California's primary being used to stoke false fraud claims. This new wave of AI-enabled misinformation intensifies fears about fabricated evidence swaying voter perception.
In response, 31 states have enacted laws aimed at curbing election deepfakes and AI-driven deception. However, these laws vary significantly in scope and formulation, with some defining prohibited conduct broadly enough to cover all AI-generated digital alterations, and others restricting only specific, malicious falsifications. This legal landscape is being tested for effectiveness and constitutionality in the upcoming midterms.
Why it matters
Legislators face a dual challenge: technology rapidly evolves while constitutional protections for political speech—especially falsehoods—are robust. Broadly framed election deepfake bans risk chilling satire, parody, and legitimate speech, as seen with California’s AB 2839, which was struck down by courts for being overbroad and insufficiently tailored.
Conversely, narrowly targeted laws like Texas’s 2019 statute focus specifically on fabricated videos intended to deceive voters within a limited pre-election period. These more precise measures are more likely to withstand judicial scrutiny and effectively address genuine harms. The central principle emerging is that laws should regulate harmful conduct rather than ban AI tools themselves, which are neutral technologies that can serve lawful purposes.
What to watch next
The midterms will serve as a litmus test for the 31 state laws against AI-related election deception. Courts will scrutinize these statutes to determine whether they strike the right balance between curbing deceptive conduct and safeguarding free expression. Laws that are too broad or technology-specific may face additional legal challenges.
Observers should also watch for legislative and enforcement approaches that adapt to evolving AI tools and misinformation tactics without prohibiting technology outright. This could include emphasizing transparency requirements, improving voter education, and promoting counterspeech as complementary strategies to legal sanctions against intentional deception.