As the 2024 US presidential and congressional elections approach, the use of artificial intelligence (AI) threatens to further undermine public confidence in election security. The AI Election Security Handbook by the Alliance for Securing Democracy (ASD) at GMF explores the dangers of malign actors’ use of AI to influence elections and the ways in which officials can counter the threat and ensure election security.

Introduction

The typical local election official in the United States is a white woman who is older than 50, earns more than $40,000, does not have a college degree, and is less likely today than in the past to have extensive experience in election administration. She is responsible for protecting her jurisdiction’s elections from a growing list of security challenges, from well-funded authoritarian regimes to conspiracy theorists, and is also continuously being asked to take on additional responsibilities in her role, often with little or no additional funding. Election officials have always been public administrators and logisticians, but after the 2000 presidential election, they were expected to be communication specialists and legal analysts. After the 2016 presidential election, they were expected to be cybersecurity experts, and during the 2020 campaign, they were even looked to for public health guidance. Now, it’s 2024, and it appears they are expected to provide expertise on artificial intelligence too.

As the 2024 US presidential cycle hits its stride, election officials face a new technological challenge amid declining public confidence in their work: artificial intelligence (AI). Trust in the US political system today is alarmingly low. Almost a third of the US electorate still falsely believes the 2020 presidential election was stolen, a belief rooted in disinformation and arguments that have no evidence to support them. The prevalence of AI technologies magnifies the challenges for maintaining trust in election integrity and democratic governance, the success of which depends on numerous dedicated actors who contribute to the health and vibrancy of democratic life—poll workers, election officials, political campaigns, the media, and the voting public. This handbook is directed specifically at election officials, who shoulder an enormous burden on behalf of the American people. Our objective is to relieve some of the burden when it comes to navigating AI in the elections space. The handbook explores how AI tools could exacerbate vulnerabilities that malign actors may exploit to undermine the integrity of the 2024 presidential election—and future election—and suggest steps for further protecting them.