Small U.S. City Tests AI to Crowdsource Civic Vision from Residents

Bowling Green, Kentucky Uses AI to Ask Residents What the City Needs—But Will Leaders Listen?

In a bold civic experiment blending machine learning and democracy, the city of Bowling Green, Kentucky has wrapped up an initiative to crowdsource ideas for its future from residents using AI. The small but rapidly growing city of 75,000 leveraged an AI-powered polling platform called Pol.is to help shape a 25-year strategic plan—raising big hopes about technology’s role in citizen engagement, and equally big questions about whether this kind of digital participation can meaningfully guide public policy.

A Fast-Growing City Without a Long-Term Plan

When Doug Gorman took office as a county leader in 2023, Bowling Green was the fastest-growing city in Kentucky and projected to double in size by 2050. Despite that growth, it lacked a cohesive long-term strategy for development. Gorman partnered with local consultant Sam Ford—who had experience with Pol.is, a machine learning platform designed to surface public consensus from large groups—to design an innovative approach to civic planning.

Ford and Gorman organized volunteers from eight key areas, including infrastructure, housing, economic development, tourism, public health, and others, to guide the process. Their idea: use AI to ask residents what they want to see in the city’s future—and listen to what they say.

The AI Tool That Turned Public Input into Patterns

In February, the team launched an online Pol.is portal, where residents could submit short ideas (140 characters or less) and vote on others’ suggestions. The system was multilingual, anonymous, and included moderation to ensure submissions came from actual residents and didn’t violate platform guidelines.

Participation exceeded expectations:

• 7,890 residents engaged with the platform (~10% of the city’s population)

• 2,000 ideas were submitted

• Over 3,900 unique suggestions were recorded

• 2,370 ideas received over 80% agreement

Google’s Jigsaw division provided additional AI analysis to highlight consensus and disagreement areas, helping to sort thousands of micro-ideas into recognizable patterns.

What the Residents Asked For

The ideas that garnered the most support were hyperlocal and practical. Among the most popular:

• Increasing local health-care specialists to reduce reliance on Nashville

• Encouraging more grocery stores and restaurants on the underserved north side

• Preserving historic buildings as the city expands

Contentious topics included legalizing recreational marijuana, expanding private education, and amending the city’s non-discrimination policies to include sexual orientation and gender identity.

According to Ford, 51% of submitted ideas were published, 31% were merged due to redundancy, and just 6% were blocked for being off-topic or inappropriate.

Can Tweets Become Policy?

Experts on democratic technology praised the initiative’s scope. Harvard’s Archon Fung called the 10% participation rate “impressive,” especially given the substantive nature of the engagement compared to a typical election.

However, many researchers stress that real democratic change depends on what happens next.

“You can’t make policy based on a tweet,” says Beth Simone Noveck, director of Northeastern University’s governance innovation lab. She notes that 140-character ideas must now be translated into actionable policy—and that public trust hinges on how city leaders handle the input.

The Limitations of AI-Driven Engagement

Others have flagged concerns about self-selection bias in digital democracy tools. James Fishkin, a Stanford political scientist known for “deliberative polling,” points out that participants in both online and in-person public forums often skew toward older, wealthier, and more educated residents. This could mean some perspectives—especially those from younger or marginalized communities—are underrepresented.

Even so, many agree that Pol.is and similar platforms offer a valuable step forward for cities seeking broad-based input without the logistical and financial burden of large-scale polling.

Will the City Act on It?

The real test of Bowling Green’s AI experiment won’t be participation metrics—it will be what elected officials do with the results. So far, the team has published the findings publicly and intends to present formal recommendations to Warren County leadership later this year.

To build trust, experts say the city will need to explain why it adopts some proposals and rejects others, making the decision-making process transparent and inclusive.

A Glimpse into the Future of Local Democracy?

As more governments explore AI-driven engagement tools, Bowling Green could become an early case study in whether AI can help bridge the gap between citizens and local government—not just in terms of gathering feedback, but also in acting on it.

The platform may have helped surface public sentiment, but true democratic innovation will depend on deliberation, responsiveness, and accountability.

Stay tuned to The Horizons Times for continuing coverage of smart cities, digital democracy, and the evolving intersection of technology and local governance.