Bowling Green, Kentucky, is pioneering a new approach to urban planning, harnessing the power of artificial intelligence to understand and incorporate citizen desires into its 25-year development plan. Faced with projections of doubling its population by 2050, the city, led by Doug Gorman, turned to the Pol.is platform, a machine learning-driven tool designed to capture and analyze public sentiment.
The project, initiated in 2023 with consultant Sam Ford, involved launching a Pol.is portal where residents could anonymously submit short (under 140 characters) ideas and vote on those submitted by others. Over a month, the platform saw engagement from 7,890 residents, resulting in 2,000 distinct proposals. Google Jigsaw’s AI then processed this data, pinpointing areas of agreement and disagreement within the community.
The AI analysis surfaced hyperlocal priorities such as expanding local healthcare to minimize trips to Nashville, attracting diverse culinary options to the city’s north side, and preserving historic architecture. Certain proposals, like recreational marijuana legalization and LGBTQ+ nondiscrimination ordinances, proved more divisive. Impressively, 2,370 out of 3,940 ideas garnered over 80% agreement, focusing on areas like stormwater infrastructure improvements and enhanced autism support services. Moderators played a role in curating the discussion, preventing repetitive or inappropriate content.
While some experts, like Harvard’s Archon Fung, commend the significant 10% participation rate, others caution against potential biases inherent in self-selected participation. Concerns about representation, particularly regarding under-representation of certain demographic groups (e.g., younger people or renters), are being raised. Stanford’s James Fishkin advocates for deliberative polling with representative samples, while jury-style citizen panels offer alternative methods.
The ultimate success of Bowling Green’s AI experiment hinges on the city’s commitment to translating these citizen-sourced ideas into actionable policies. As Beth Simone Noveck of Northeastern University emphasizes, these 140-character ideas serve as a starting point, requiring further development and careful consideration. Transparency in the decision-making process, detailing which ideas are adopted and why, is crucial for maintaining public trust. The project results have been published, and recommendations will be presented to county leaders later this year.
Photo by David Skyrius on Pexels
Photos provided by Pexels