LWVUS Submits Statement to Senate around Hearing on AI and Deepfakes in Elections

LWVUS submitted a statement for the record to the Subcommittee on Privacy, Technology, and the Law in the US Senate. The Hearing was titled Oversight of AI: Election Deepfakes.

Statement for the League of Women Voters of the United States 

US Senate Subcommittee on Privacy, Technology, and the Law 

April 16, 2024 

On behalf of our 500,000 members and supporters, the League of Women Voters of the United States (LWVUS) urges the Committee to take action to curb the negative impacts of Artificial Intelligence (AI) in our elections, including by combatting deepfakes. Deliberately false AI-generated political content often attempts to garner support for or opposition to particular candidates or cause confusion around a person’s voting time, place, or manner. The League is becoming increasingly concerned that AI-generated disinformation will undermine the role of voters and corrupt the election process.  

LWVUS was founded in 1920, working on the front lines of voter education to assist newly enfranchised women in casting their ballots following the ratification of the 19th Amendment. For over a century, LWVUS has remained committed to our mission to empower voters and defend democracy. The League focuses on advocacy, education, litigation, and organizing with our grassroots network of more than a half-million members and supporters across over 750 Leagues in all fifty states and the District of Columbia. The League is nonpartisan — neither supporting nor opposing candidates or political parties at any level of government — and is committed to protecting the freedom to vote. 

The League derives our policy positions based on grassroots member support and consensus. As stated in our position on a citizen’s right to know and citizen participation:  

The League of Women Voters of the United States believes that democratic government depends upon informed and active participation at all levels of government. The League further believes that governmental bodies must protect the citizen’s right to know by giving adequate notice of proposed actions, holding open meetings, and making public records accessible. 

Additionally, as stated in our position on campaign finance: 

The League of Women Voters of the United States believes that the methods of financing political campaigns should provide voters sufficient information about candidates and campaign issues to make informed choices; [and] ensure transparency and the public’s right to know who is using money to influence elections.  

These positions are applicable to the issue of deceptive AI campaign communications and deepfakes. Voters deserve access to true, genuine, and complete information about elections and the candidates seeking their votes. The distribution of disinformation, especially online, has been used in recent elections to sow polarization and distrust in election results in our country. It is crucial that we address the avenues of mis- and disinformation that circulate around an election.  

As an organization dedicated to empowering voters, we work to simplify the voting process and make voting accessible, breaking down barriers to participation. Our Democracy Truth Project aims to strengthen democracy and restore trust in the electoral process by combating mis- and disinformation. Through our Democracy Truth Project, Leagues have been trained to engage and report unusual social media and AI-supported election-based misinformation through our partnership with the Algorithmic Transparency Institute (ATI) to expand the Civic Listening Corps (CLC).  

The CLC enables the League to contribute to a national mis- and dis-information data set, providing trends across organizations. It allows for state and local Leagues to report to our national office to ensure we are looking at and identifying trends of mis- and disinformation and how best to rebut that information.  

Our partnership with ATI, part of the National Conference on Citizenship, combines their technical expertise with the League’s ‘people power’ to identify, analyze, report back, and take action on the latest disinformation trends.  

In its first year (2023), the Democracy Truth Project cohort hosted over three hundred meetings with election officials, forty-seven training courses on mis- and disinformation, and 265 events focused on the election process. 

By collaborating with stakeholders across sectors and harnessing the power of litigation, communications, and grassroots organizing, we aim to fortify our democratic institutions against the emerging threats posed by AI-driven disinformation campaigns. However, it should not fall solely to organizations such as the League to provide verified, reputable information to voters and ensure transparency in our election process. Congress should similarly break down barriers to participation by reducing the influx of election mis- and disinformation.  

The 2024 election is just months away, and primary season revealed some of the deepfake tactics being used to manipulate voters. Two days before the New Hampshire presidential primary, robocalls were sent to New Hampshire voters with a deepfake, simulated voice of President Joe Biden discouraging them from participating in the primary. The New Hampshire robocalls urged recipients not to vote in the primary and to “save” their vote for the November 2024 US Presidential Election.  

In response, the League of Women Voters of New Hampshire, the League of Women Voters of the United States, and individual New Hampshire voters filed a federal lawsuit against Steve Kramer, Lingo Telecom, LLC, and Life Corporation for voter intimidation, coercion, and deception ahead of the presidential primary. This litigation underscores the critical need for proactive measures against malicious actors seeking to undermine the integrity of our democratic processes. Bad actors’ utilization of AI-generated content impersonating political figures constitutes a grave threat to voter confidence and participation. 

The example in New Hampshire is one of many instances of AI-generated disinformation this year. Fake and deceptive content has been used to target voters of all political party affiliations and has been condemned as unfair and undemocratic by elected leaders across the political spectrum. AI companies have said their tools should not be used in political campaigns, but enforcement has been spotty. The incomplete framework of regulation on AI leaves us vulnerable to efforts to undermine our elections — both from domestic and foreign entities. It is imperative that Congress curb bad actors’ ability to use AI in our elections.  

Voters deserve to know that the political advertisements they see and hear are free of misleading information or fraudulent misrepresentation. The League hopes to be a resource and partner in this endeavor to ensure that our elections are free and fair and that voters can make informed voting decisions. If you have questions or would like to discuss this further, please contact Jessica Jones Capparell, Director of Government Affairs at [email protected].