The first-ever AI + Elections Hackathon

On February 26th, 2026, the Mechanics of Democracy Lab hosted a joint day of programming - combining the AI + Elections Clinic Bootcamp and the first-ever AI + Elections Hackathon. 

The bootcamp brought together over 40 election officials and experts from across the state to collaborate on a hands-on learning experience that explored and tested new AI tools, AI-integrated workflows, and new tech parameters in an experiential learning environment with real-life scenario activities. 

Between learning how to navigate new AI tools and applying them to carefully crafted scenarios, thanks to the team at The Elections Group, elections officials were given a different kind of exercise–one that tested their AI capabilities with the help of ASU students. 

Led by MODL’s Director of Curriculum Development, Dr. Allison JoAnn Lester, the AI + Elections Hackathon brought together over 40 students and elections officials to tackle one central question:

How can AI responsibly support and strengthen election administration?

Rather than designing solutions in isolation, students worked directly with election officials to understand the real operational challenges facing election offices. Participants listened closely, asked questions, and then began building tools grounded in practical needs—from improving accessibility of election materials to helping offices respond to voter inquiries and monitor misinformation.

After thorough development and deliberation, a panel of judges – Scott Jarrett, Maricopa County, Dana Lewis, Pinal County, James Heiman, Pinal County, Lester Godsey, ASU, and Dr. Jennifer Miller, ASU – declared the winning teams. 

In first place, Kirtankumar Thummar and Shubham Tiwari developed an AI engine that converts scanned and handwritten election documents into digitally readable, ADA-accessible formats. Using Google Vertex AI and Document AI, the system performs image-to-text conversion, enriches documents into structured data, and integrates human confirmation directly into the workflow. The tool is designed to reduce repetitive administrative workload while improving accessibility and long-term usability of election documents.

In second place, Dvir Hamu and Nathan Sheyman built a secure, multilingual AI assistant that helps election offices respond to voter questions quickly and clearly. The tool delivers jurisdiction-specific, source-verified information via SMS, email, or chat. It translates complex procedures into plain language and uses only official government data—without collecting sensitive personal information.

In third place, Gnana Surya Teja Reddy, Shreesha Reddy Eswara Reddy, Avantika Suresh, and Allan Jerrold Kirubakaran designed an AI system that analyzes rapidly spreading election misinformation. The platform allows voters to submit questionable content for credibility checks while alerting election officials to emerging misinformation trends. By using large language models to analyze online content and surface sources, the tool helps election administrators detect and respond to narratives before they spread widely.

For many of the students, the hackathon was a great opportunity to work directly with election administrators. For election officials, it was an opportunity to get a glimpse of how emerging technologies could support the future of election administration.