In the course of writing this article, we have spoken to numerous election officials and security experts about their apprehensions and ambitions regarding generative artificial intelligence. Within the elections community, many officials have expressed grave concerns about what generative AI might mean for election security. That sentiment aligns with recent media discourse highlighting the dangers posed by AI. An illustrative May 2023 article in the Washington Post described an increase in phishing attacks attributed to AI, noting that such attacks are “just the beginning . . . as attackers use artificial intelligence to write software that can break into . . . networks in novel ways, change appearance and functionality to beat detection, and smuggle data back out through processes that appear normal.”
Other commentaries recount the ways that AI may make it easier to impersonate election officials, offices, and system vendors in order to infiltrate election infrastructure and mislead the public. Meanwhile, some election officials worry that excessive emphasis on AI’s frightening capabilities itself breeds more peril. David Triplett, the elections manager for Ramsey County, Minnesota, wonders if “all this talk about AI’s power is fueling even more paranoia and confusion about the trustworthiness of elections.” Such concerns seem valid given recent sensationalist narratives stoking public fears to propagate baseless claims of AI-driven election rigging.
How much should election offices and vendors worry about the security threats posed by advances in AI, and what steps should they take to prepare for them? This article is the inaugural piece in a series delving into AI’s potential effects on American democracy. It examines generative AI’s capacity to disrupt the security of election offices and election system vendors. Specifically, we look here at how AI changes (and does not change) the cybersecurity situation for election offices and election infrastructure. We also look at the threat that generative AI poses to the ability of election offices to function as authoritative sources of election information and records. We then detail measures that government, the private sector, and the media must take to guard against AI’s risks and build public confidence in the integrity of election outcomes in the AI age. Later pieces in this series will consider, among other issues, AI’s effects on manipulated media in political advertising, election administration, voter suppression efforts, and the public comment process.