The conversation about artificial intelligence of late tends to center on national security, but the intersection of generative AI and social media is a more immediate concern for state and local election officials. The January 2024 AI-generated robocall that imitated President Biden and targeted New Hampshire primary voters showed how fast a speculative threat can become operational.
The Brennan Center for Justice, the Bipartisan Policy Center (BPC), and Protect Democracy have all studied this problem. Generative AI is being used to exploit the social media feeds of voters and local election workers, largely in three ways:
- Automated Harassment at Scale: Harassment campaigns used to require a large number of people coordinating fake or misleading accounts. AI is removing that requirement by lowering the barriers to entry. A single actor, or even an automated system, can generate thousands of unique, targeted comments that name a specific clerk, push the attack into trending feeds, and overwhelm staff with threats and doxxing (posting someone’s home address or personal information to encourage further harassment).
- Impersonation of Officials: Bad actors scrape public materials, including video and audio footage of officials, to create realistic videos or voice clones that they post on social platforms. These clips might falsely announce changes to voting procedures or conditions, such as precinct closures or machine failures, potentially sparking chaos at polling places or confusing voters before election officials can issue a correction.
- Images of Misconduct That Never Happened: Tools like ChatGPT and Google Gemini can produce photorealistic images of misconduct that never happened, such as poll workers burning ballots. Once those images are seeded on visual platforms like Instagram or Reddit, the initial shock drives rapid sharing, leaving a lasting impression of misconduct even after the content is flagged as fake.
Why Election Officials Should Take Note
Running an election today is inseparable from what voters see about it online. The Brennan Center’s 2026 Local Election Officials Survey found that more than half of respondents are worried about the safety of their staff, and nearly a third reported experiencing threats, harassment, or abuse.
It is important to remember that AI does not directly create these pressures; it makes existing risks cheaper and easier to amplify. Social media platforms reward speed, emotion, and visual engagement, which are things generative AI does fairly well and continues to improve over time. Attackers can cause significant harm by injecting false information into the social media feeds people already rely on, shaping public perception and undermining trust. Their intent is to steer enough people to false information by turning platforms into megaphones that question official procedures and security measures.
Proactive Actions
- Move your official website to a verified .gov domain and claim verified accounts on the social platforms your voters use. Doing this early helps voters know where to find real updates and information.
- Expand incident response and continuity plans to cover targeted harassment, not just technical outages. Train staff to lock down personal social media accounts and report doxxing or sudden waves of threats to law enforcement.
- Build direct relationships with community leaders, civic organizations, and advocacy groups you trust. They will often hear a rumor moving through private group chats before you do and can help share accurate information through those same channels.
- Tell your story of secure processes and testing protocols before election day, and include high-quality videos and photographs where possible, because people often engage with visual content more than plain text. When voters have already seen how your process actually works, a fake claim has less room to take hold. This is called pre-bunking: getting the truth in front of people first.
As AI capabilities advance, so does the risk of voter manipulation through false information. For details on verifying AI content and digital provenance, review the Resource Library section of this newsletter.
The Situation Room focuses on real security incidents and threats in the news relevant to election security. To review previous issues, see the newsletter archive.