As the specter of misinformation looms large over upcoming elections, at least seven states are gearing up to introduce legislation in 2024 to restrict the use of artificial intelligence (AI) in political campaigns. This move comes in response to concerns raised by experts about the potential deluge of misinformation that could impact electoral processes.
Urgent Call for State Action
Robert Weissman, president of the election advocacy group Public Citizen, emphasizes the urgent need for state action. He stresses the importance of establishing standards that deem deepfakes impermissible in the context of elections. Deepfakes are digitally altered images or videos that misrepresent the reality of a person or related events.
Proposed Bills Take Aim at Deepfakes
Pending bills in the seven states, including New York, Florida, and Wisconsin, seek to ban AI-generated deepfakes effectively. These bills propose mandatory disclosures when AI technology creates images, videos, or audio of political candidates. If enacted, these restrictions would impact roughly half of the U.S. electorate.
States in Focus
All states with pending bills, including New Hampshire, South Carolina, and New Jersey, have competitive congressional races in the upcoming year. Meanwhile, five states—California, Texas, Michigan, Washington, and Minnesota—already have laws addressing deepfakes, with California considering further measures, potentially including a complete ban on AI in political communications.
Federal vs. State Initiatives
While states are taking proactive steps, Congress at the federal level lags in addressing AI-related issues. The Federal Election Commission is contemplating rule changes prohibiting federal candidates from using generative AI tools to misrepresent their opponents. State legislatures, however, could take broader action by banning such deceptive practices at all levels of elections.
Varied Approaches to State Legislation
Bills across the seven states exhibit variations in their approach to addressing AI-made deepfakes. While all seek to compel content creators to disclose the use of AI technology in election-related materials, the specifics differ.
A year-round deepfake law is in effect in Washington, but Florida, Wisconsin, and New York are considering similar provisions that would extend this coverage. Other states, such as New Hampshire, South Carolina, and New Jersey, propose measures applicable only in the 90 days leading up to an election, with Illinois suggesting a 30-day limit.
Who’s Covered and How Enforcement Works
A key divergence among the bills lies in defining who falls under their purview and how enforcement will occur. Wisconsin’s bill targets candidates, parties, and political committees, while others extend to anyone distributing election materials using generative AI. Some accounts, including those in South Carolina and New Hampshire, also encompass corporations.
The New Jersey bill introduces the possibility of “registered voters” suing for injunctive relief, joining political candidates in legal action. Meanwhile, New York, Florida, Illinois, and Wisconsin are considering delegating enforcement responsibilities to government bodies.
Potential Legal Challenges Ahead
As these state laws targeting AI in political campaigns multiply, election attorney Hannah Miller of Wiley Rein LLP in Washington anticipates likely court challenges. The constitutionality of these laws remains uncertain, and Miller notes that the coming year may witness legal battles over their validity. The evolving landscape underscores the delicate balance between protecting electoral processes and navigating the legal complexities surrounding AI use in politics.