Democracy Works and AI

Helping voters navigate the first elections in the generative AI era

Since 2010, Democracy Works has helped millions of Americans vote by providing them with reliable, useful voting guidance through the sites, apps, and organizations they trust most. 

Our experts research voting details and guidance for thousands of elections every year from authoritative government sources and regularly review guidance to ensure accuracy.

Just after the 2022 midterm elections, OpenAI released ChatGPT, a chatbot that allows users to interact with its powerful large language model (LLM). This milestone made it clear that the November 2024 elections will be the first general election of the generative AI era. 

While this and many subsequent advancements in AI present an opportunity to introduce novel ways to support voters, these powerful new LLMs also introduce real risks that threaten to erode trust in our electoral system. 

Generative AI systems have impressive text-prediction abilities but often lack the sophistication to ensure text is accurate, comprehensive or timely. Because misrepresenting even the smallest detail about elections can have severe consequences, implementing generative AI in any experience where users may seek or encounter voting guidance risks disenfranchising voters.

To safeguard voters in the first elections of the generative AI era, Democracy Works is working hard to ensure that all sites and apps powered by our comprehensive election data—including TurboVote and third-party experiences—are reliable and trustworthy.