Navigating the intricacies of state Capitol legislation can often feel deliberately opaque and confusing to the average citizen. To bridge this gap between legislative actions and constituent understanding, a new artificial intelligence (AI) application called CivicLoon has been developed by local programmer Colin Lee. This app aims to demystify the legislative process, making it more accessible and transparent.

Colin Lee, a Lakeville-area software engineer and principal mobile architect for a Texas-based AI company, conceived the idea for CivicLoon years before he began coding. His motivation stemmed from personal experience as a DFL candidate who ran multiple times for a state House seat. During his campaigns, Lee was repeatedly struck by the profound lack of voter awareness regarding candidates and policies on their ballots, noting instances where voters misgendered his opponents and demonstrated a general lack of understanding. Despite never winning a seat, Lee continued to run on principle, believing that challengers serve a crucial purpose in forcing conversations on substantive issues like healthcare, which he felt other candidates often sidestepped in favor of hyperlocal concerns.

CivicLoon functions by pulling in bill text, news coverage, and committee schedules directly from the Minnesota Legislature. It then processes this information and presents it to users in clear, plain-language summaries. A significant feature of the app is its extensive language support, currently offering summaries in 30 languages. However, Lee acknowledges that the quality of translation can vary, particularly for languages with limited online source material, such as S’gaw Karen, which is spoken by a large Karen community in St. Paul.

A distinctive technical aspect of CivicLoon is its commitment to user privacy and reliability. Unlike many applications that rely on remote servers for AI processing, CivicLoon runs its AI model directly on the user’s phone. This design choice, which Lee has also applied professionally in developing AI chatbots that keep sensitive data off third-party servers, ensures that the app has “no cloud reliance” and “only depends on itself,” thereby enhancing data privacy.

This approach stands in stark contrast to existing resources, such as the bill summaries available on the House and Senate websites. These official summaries are often brief, written in complex legal jargon (legalese), and are exclusively in English, severely limiting their accessibility to a broader public. CivicLoon’s plain-language, multi-lingual summaries offer a much-needed alternative.

University of Minnesota law professor Daniel Schwarcz, who specializes in AI and the legal system, weighed in on the capabilities of AI in this context. While legitimate concerns about AI bias and hallucination (when models generate plausible but incorrect responses) persist in many applications, Schwarcz notes that summarizing and translating documents are areas where AI models generally excel. He stated there is