How AI Can Help (and Hurt) Election Officials

Why It Matters

The rapid AI escalation could amplify election interference and erode public trust, but proactive adoption of AI safeguards can improve efficiency and resilience for vulnerable jurisdictions.

Key Takeaways

• AI capability 675× stronger than in 2020, intensifying election risks
• Autonomous AI agents can create thousands of false posts, deepfakes, DDoS attacks
• Only 16% of election offices use AI; ~50% request guidance
• ASU’s AI‑Elections Clinic offers boot camps, prompt libraries, and risk‑management tools
• AI can boost turnout forecasts, streamline training, yet human oversight stays essential

Pulse Analysis

The past few years have witnessed an exponential rise in generative AI, with large language models improving roughly every seven months since 2019. That acceleration translates to tools that are hundreds of times more capable than those available during the 2020 U.S. election cycle. For election officials, the implication is clear: AI can now produce persuasive political content, fabricate realistic videos, and even conduct autonomous reconnaissance at scale. The sheer volume and speed of these outputs lower the entry threshold for foreign or domestic actors seeking to sway public opinion, turning misinformation campaigns into a near‑industrial process.

Beyond the obvious propaganda risks, AI introduces technical vulnerabilities that could directly disrupt election infrastructure. Autonomous agents can flood official websites with traffic, creating denial‑of‑service attacks that cripple voter information portals at critical moments. Deepfake videos and AI‑generated articles further muddy the information environment, making it harder for voters to distinguish fact from fabrication. A Brennan Center survey underscores the anxiety in the field—only 16% of jurisdictions have deployed AI tools, yet nearly half are actively looking for guidance to mitigate these emerging threats.

Recognizing both danger and opportunity, a coalition of nonprofits, academia, and election‑security groups has launched resources to help officials navigate the AI landscape. The AI and Elections Clinic at Arizona State University, led by former Maricopa County supervisor Bill Gates, offers boot camps, curated prompt libraries, and risk‑management playbooks. Practical use cases already show promise: AI‑driven turnout models inform staffing decisions, while chat‑based assistants summarize complex regulations for poll workers. The consensus among experts is that catastrophic AI interference in the 2026 midterms is unlikely, but the early adoption of safeguards and human‑in‑the‑loop oversight will be decisive in preserving election integrity.