Kentucky counties conducted a month-long “town hall” with nearly 8,000 residents attending earlier this year, thanks to artificial intelligence technology.
Bowling Green, the third largest city in Kentucky and part of Warren County, faces a massive population spike by 2050. To prepare for this, county officials wanted to incorporate community input.
Community outreach is a tough business. Town Halls are widely employed, but do not tend to attract large crowds. When people come, it is a self-selective pool of people with only strong negative opinions and does not represent the entire town.
Meanwhile, gathering the opinions of most cities through online surveys results in a large dataset that makes it difficult for staff and volunteers to look into it and understand its meaning.
Instead, Bowling Green county officials have the AI do the part. The participation was large. In an online survey for about a month, around 10% of Bowling Green residents expressed their opinions on policy changes they would like to see in the city. The results were then integrated by AI tools and became policy reports, which are still available to the public to see on their website.
“If there's a city hall meeting on these topics, 23 people will appear,” Warren County Judge Doug Gorman told PBS News Hour in an interview published this week. “And what we just went to was America's largest city hall.”
Bowling Green Experiment
The county, with the help of a local strategy company, launched its website in February where residents could submit anonymous ideas. Their research using Pol.is is an open source online polling platform used around the world for citizen involvement, and has been particularly successful in Taiwan.
The prompt was open-ended and asked participants what they wanted to see in the community over the next 25 years. I was then able to continue to participate by voting for other answers.
Over the 33 days when the website accepts responses, around 8,000 residents weighed over a million times, sharing about 4,000 unique ideas for a new museum, expanding pedestrian infrastructure, green spaces and more.
The answers were then compiled into reports using Sensemaker, an AI tool from Google's Tech Incubator Jigsaw.
Finally, Sensemaker found 2,370 ideas that at least 80% of respondents agree. Some of the most agreed ideas included increasing the amount of medical professionals in the city, so residents didn't need to resort to services an hour away from Nashville, reuse empty retail space, or add restaurants to the north side of the city.
Online surveys have allowed people the county to reach, such as those who were politically freed and those who could not find time to attend City Hall from work.
This format was also excellent at reaching immigration by providing surveys in multiple languages and automatically translating responses. It was welcomed by people like Daniel Tarnagda, an immigrant from Burkina Faso, and the founder of a local nonprofit who leads a soccer team for immigrants under the age of 18 who are struggling to speak English.
“I knew people wanted to be part of something, but if you don't ask, you don't know,” Tarnagda told PBS.
Volunteers on the project will now compile ideas from the report and make specific policy recommendations to county leaders by the end of the year. According to a survey conducted by Jigsaw with local leaders, AI saved the county an average of 28 business days.
Agreeing beyond the party line
The Bowling Green experiment was the sensemaker's first large-scale proof of concept, Jigsaw wrote in a blog post earlier this year.
One of the most impressive things I found on Bowling Green is that when the idea was stripped of its political identity anonymously, its constituents agreed to a lot.
“When most of us don't take part, the people who do it are usually the people with the strongest opinions, perhaps the most information and perhaps the most angry, and you start to have caricature ideas about what the other side thinks and believes.
Jigsaw announced this week that it is partnering with the Napolitan Institute, a research and public voting agency founded by renowned polling officer Scott Rasmussen. Unlike the bowling green experiment, the goal is not to policy, but to understand where the country stands.
The Possibility of AI: Good and Evil
There are still concerns that are essentially linked to this experiment with AI in local governance. The Bowling Green Research website explicitly points out that “personal information has not been captured and demographic data is not stored,” but this does not necessarily mean that future applications of this will follow.
Artificial intelligence is the subject of privacy concerns due to its vulnerability to data breaches, and becomes a problem when people obtain DOXX for political beliefs they have confidently submitted.
AI also has a problem with the bias that the creator is burned into an algorithm. Last month, researchers discovered that Elon Musk's Grok Chatbot consults Musk's own (surprisingly controversial) opinions before answering sensitive questions. Such flaws will be insurmountable if AI generates neutral policy proposals.
However, if these concerns are addressed properly, AI could completely revolutionize citizen engagement. It also shows a pathway to migrating past political polarizations and enacting tangible changes, similar to how AI created space for divided communities of bowling greens to find common positions.
