CivicLoon is aimed at making Minnesota legislation easier for residents to follow as lawmakers trim the set of bills that could advance during the final month of the session, according to the app’s developer and other experts. Colin Lee, a software engineer who built the project, said the state’s legislative process can feel designed to confuse people who are not familiar with how bills are written and advanced.
Lee said that gap between what elected representatives are doing and what constituents can see and understand was a recurring frustration for him as he ran as a DFL candidate for a state House seat covering parts of Dakota and Scott counties. In comments about his experience trying to talk to voters, he described encountering misunderstandings about the candidates and the choices people planned to make.
“As an ordinary person, it doesn’t feel like you really have a voice. Lobbyists are there every single day of every work week the session is open. As a normal person, you might go there once in 10 years,” Lee said. He also said that during past campaigns, he was repeatedly struck by limited voter knowledge about the people and policies on ballots, including confusion he described in conversations about the race.
CivicLoon, Lee said, pulls in bill text, news coverage and committee schedules from the Minnesota Legislature and then presents that material to users in plain-language summaries. Lee said the app currently supports 30 languages, while acknowledging that translation quality can vary depending on the amount of source material available in a given language online—highlighting S’gaw Karen, which he flagged as a potential weak point. He described work on improvements, including the ability to track and receive updates on specific bills and broaden translation coverage across the app.
One of CivicLoon’s technical features, Lee said, is that it runs the AI model directly on the user’s phone rather than sending data to remote servers. “It has no cloud reliance,” he said. “It only depends on itself.” Lee said that design choice was driven by privacy and reliability considerations.
In addition to summarizing and translating documents, the app includes a feature intended to assess a bill’s chance of passage, but Lee said that part remains rough. He said one reason is that the AI model is working from legislative text that can include positive or promotional language added by lawmakers, which may complicate efforts to gauge momentum. Lee said he plans to add other signals over time, including analysis of legislators’ newsletters and public statements, which he said could eventually support more reliable predictions about which legislation has genuine momentum.
Daniel Schwarcz, a University of Minnesota law professor who studies AI and the legal system, said while bias and hallucinations are legitimate concerns in many AI applications, summarizing and translating text is comparatively stronger. “There’s a lot of research that one context in which AI is pretty darn good is in summarizing text that you give it,” Schwarcz said. He said he would have “comparatively less worry” about bias where AI is summarizing a text input than in other applications, such as recidivism-prediction algorithms used by judges, where training data reflecting historical inequities can compound into discriminatory outcomes.
Schwarcz also framed the question as whether the public will use CivicLoon enough to make a difference, noting that AI-driven summaries from other products—such as Google’s AI Overview—are available to many users. Lee, meanwhile, described CivicLoon as an attempt to make bill information accessible beyond English legalese and beyond the brief summaries that exist on House and Senate websites.
The League of Women Voters Minnesota’s Amy Perla said the organization has not used the app but supported the concept of tools that increase public participation in politics. “Any way we can increase civic engagement and civic education, we think is worth exploring,” Perla said.