Table of Contents
State Tech Pulse Policy Brief: AI-Generated Content Influence on Elections
Background: Primaries for state and federal elections kicked off this month, and states have been busy filling in gaps to address election misinformation. Momentum started in 2023, with golden domes across the US looking to address concerns about widespread AI-driven misinformation ahead of the Presidential election, with states continuing to act. This policy brief explores the 5 W’s (Who, What, Where, When, Why) around oversight of AI-generated content used to influence elections.
Public Policy Problem: AI-generated election content has already had an impact and there are no federal protections to oversee these activities. Recently, AI-generated content in LA’s mayoral race has raised questions and concerns similar to those raised during the last Presidential election cycle, when widespread AI-generated robocalls mimicking a candidate were uncovered. Election AI-generated content takes many forms, from realistic audio to videos and pictures.
Primaries for elections across the US have started. In an election year, primaries begin in May and end in August to determine who will be listed on the final ballot on election day.
Proposed Policy Solutions:
Many questions arise about how AI can be used to influence elections and what should be allowed in election-related AI-generated content. Although there have been long-standing state-election laws aimed at transparency and insight into who is influencing campaigns, the spread of political speech in other spaces, like social media, has raised new questions about whether old legal frameworks are responsive to the “new public square.”
Evidence in state legislation on this topic this year shows how legislators are grappling with this new dynamic. New legal definitions in proposals and new laws include:
“synthetic media”
“artificial intelligence”
“materially deceptive”
“deepfake”
“election misinformation and election disinformation “
Many of these new laws are being tested for the first time. For example, in California, multiple laws have been enacted regarding AI-generated content in elections, and some have been halted by the courts. One law required the removal or labeling of “deceptive” content and the other law prohibited political advertising including “deceptive AI-generated” content close to elections. Both laws were struck down for being in conflict with the First Amendment. The other law (AB 2355) that is currently active requires disclosures of AI-generated content in political advertisements.
New laws related to restricting content have faced legal challenges, whereas laws focused on transparency and disclosures are more viable.
Who and When? Since 2023, state legislators across the US have tightened and built on election code laws traditionally focused on misrepresenting election-related details or candidates and disclosures related to who funded political ads.
What, Where & Why?
Although more than half of the states have taken action on this topic over the past few years, others continue to prioritize it. So far, four state governors have signed new AI-election laws.
Maryland’s General Assembly’s adjourned last month after debating multiple measures related to AI-generated content used during an election. This week, the Governor signed SB 141, which will go into effect next month. The proposals in Maryland take new law different approaches than other state-led efforts. First, the new law prohibits a person from creating a deepfake to push false information to influence an election. The law also allows the State Board of Elections to step in to require the correction or removal of election misinformation or disinformation. Another measure (SB 697) that lost momentum during the legislative session would have required social media platforms to disclose activity related to the platform amplifying specific political content over others and prohibiting a non-US citizen from using an AI bot to publish, distribute, or disseminate campaign material or artificially amplify political content.
Last month, three Governors signed new laws requiring disclosure of AI use in elections.
In four other states, AI-election-related laws are continuing to move through the process.
Louisiana has a few more weeks left for its legislative session. A bill (HB 459) focused on requiring a disclosure in electioneering communication a few steps away from the finish line. According to the bill’s intent, the Louisiana State Legislature indicated this measure is needed to protect the electoral process and prevent the public from being misled by an image or likeness of a candidate created using artificial intelligence. Another bill (HB 639), one vote away from passing this legislative session, focuses specifically on transparency for AI-generated political robocalls, requiring disclosure that the content is AI-generated. A bill (HB 734) no longer in play took a different approach by creating an “AI Bill of Rights,” with one of the 9 rights being the right to know if political campaigns include AI.
Measures being debated on this topic in six other states did not make it.
During this year’s Alaska Legislative session, state legislators debated an (SB 64) election omnibus bill focused on changes to election administration, which the Governor ended up vetoing. The original version of the legislation prohibits the use of synthetic media, including AI, to influence an election without disclosing its’ use. The law would have allowed an individual to sue if they were harmed and request that the content be removed. Legislators in the House took this out.
Over the past two years, the Georgia General Assembly debated legislation (SB 9) around allowable and prohibited uses of AI. The Senate version of the bill focused on prohibiting youth access to sexually explicit AI-generated content, and the House changed it to focus on AI-generated content in elections. Ultimately, the legislators could not agree, and the bill died.
Right at the finish line, Virginia’s bill (SB 141) to require disclosure of AI-generated content in electioneering communications failed to pass.
Connecticut’s bill, HB 5342, would have required a disclosure of synthetic media use in elections and failed to pass.
Do you have leads, tips, corrections, feedback or resources you would like to share? Send your advice to [email protected].
Disclosure: This is a human-written and driven publication. As a small business owner and mighty team of one, I use AI tools to optimize my small business operations as a part of my admin tech stack. Regarding this publication, AI is mainly used to help with catchy titles, as a thesaurus when writing and a partner when creating cartoons. (Thanks, Canva, and not an ad!) As a secret doodler, I add my human touch using my digital pad and pen. I also use Grammarly, with AI built in, to help with copy editing/grammar (again, mighty team of one!) Thanks for reading. 😊
