WASHINGTON D.C. – The Alliance for Securing Democracy at the German Marshall Fund of the United States released “The ASD AI Election Security Handbook” as a resource for elections officials responding to unprecedented threats to election security supercharged by technologies like artificial intelligence (AI). 

Crafted by leading experts in emerging technologies and election integrity, the handbook describes the new threats looming over this election cycle and offers real solutions for officials to safeguard democratic processes. From AI-generated disinformation campaigns to sophisticated cyberattacks targeting election infrastructure, the authors analyze the ways these technologies might be misused, share what experts are watching out for this cycle, and offer specific action items for officials and policymakers to safeguard election security. 

ASD Senior Fellow for Emerging Technologies Lindsay Gorman said: “Generative AI is upending our relationship to information, democratizing the ability to create realistic fake calls, videos, and images. Nowhere is quality information more vital to a democracy than in selecting political leaders, and now AI and the largest election year in history are headed for a near-on collision. Equipping front-line election officials with the tools and tactics to meeting evolving technology threats is an urgent national priority.”

“The emergence of generative AI tools comes at a time when the demands and expectations facing election officials continue to grow, trust in US elections is at an all-time low, and threats to elections continue to increase,” said ASD Senior Fellow for Elections Integrity David Levine. “Election officials are the biggest reason recent US elections have been successful, and this handbook offers tips that should help them replicate that success in 2024.”

These are some of the steps the authors recommend that election officials take:

  • Simulate AI threats to election infrastructure. Conducting mock elections and tabletop exercises to test resilience against AI-driven phishing campaigns and disinformation help officials identify vulnerabilities in voter registration databases, voting systems, and election day operations, testing partnerships with state, federal, and academic entities for comprehensive threat assessment.
  • Maintain hard copies of voter registration for resilience against ransomware. Keeping organized, easily accessible hard copies of voter registration cards to mitigate the impact of AI-enhanced ransomware attacks. This ensures continuity in the signature verification processes for mail ballots, even in the event of digital system compromises, safeguarding against delays and exploitation. 
  • Pilot content authenticity technologies before the election. Exploring technologies that embed data on content creation and modification to combat deepfakes, enhancing trust in digital communications. Some election officials—particularly those in well-funded, tech-savvy jurisdictions—should pilot integrating these technologies into their election infrastructure to increase public trust.
  • Create reading rooms for public records requests. Mitigate the administrative strain caused by a potential surge in AI-generated public records requests by expanding public records reading rooms to proactively offer frequently requested public records streamlines the process and alleviates the burden on election officials.

 

Read the ASD AI Election Security Handbook here.

The views expressed in GMF publications and commentary are the views of the author alone.