Publication select ▼
MADISON HEIGHTS — As artificial intelligence continues to develop, society is grappling with the best ways to use it. This includes the city of Madison Heights, where officials say AI has potential to do good — if handled responsibly.
To this end, the city’s Information Technology Advisory Committee is currently drafting guidelines for how to leverage AI safely and ethically at City Hall.
“We need to establish what our administrative departments should use it for, but also the things that should never be done, like putting sensitive information on the public record in an LLM (large language model) like ChatGPT. That would be a risk,” said Mark Bliss, the mayor pro tem of Madison Heights and founder of ITAC. “I also think when AI systems are used anywhere but especially in government, people need assurance that the decisions that affect their lives are fair, transparent and explainable, which means it must have human oversight and accountability.”
Bliss works as vice president of marketing at a firm that specializes in AI implementation for mid-market companies that lack technology infrastructure.
“It’s a topic I’m very passionate about. I think the main thing with AI is it’s already here, and the ultra-large companies and cities are already leveraging it to maximize efficiency, but it’s the smaller companies and cities that could benefit the most, since they simply don’t have the staff and funding,” he said. “If they use AI to make each person on their staff even 5% more efficient, that’s way more powerful for a smaller city than a larger one.”
AI models are trained on data sets that are fed to them and perform tasks when given prompts by the user. A well-trained AI system can cross-reference information in the blink of an eye, creating summaries that answer questions. AI can also be used to add subtitles to broadcasts, translate languages in real time, create convincing audio and visuals, and more.
“One use for city government is generating minutes for a public meeting that was held,” Bliss said. “In the old way, this would be done where the whole recording must be gone through, and the minutes are taken shorthand during the meeting and then amended and finalized afterward. And it must be done, since every public meeting must have filed minutes. But with AI, you can cut out 90% of the human effort. The AI simply takes the transcript and creates minutes off a standard template, and then a staff member re-reads it to make sure it aligns with what happened. It saves a lot of time.”
Similarly, AI can be used to quickly summarize documents that are hundreds of pages long. Bliss said this can be particularly helpful when evaluating multiple bids in a short period of time.
He noted that other cities have used AI chatbots as virtual assistants on their websites, answering frequently asked questions, communicating with guests about city services, and sharing information in a spoken voice that is easy to understand.
Some cities have also used AI as a form of financial oversight, Bliss said, where the AI can flag unusual transactions and spending patterns. And then there is the way AI can quickly draft job descriptions, internal memos and policy language that cities use on a regular basis.
“Another great use is real-time audio transcription,” Bliss said. “An AI can add subtitles to a meeting broadcast in real time and even translate it into other languages so more residents can follow along. Making government services available to more diverse communities like ours in Madison Heights are incredible use cases, but you must do it right.”
Indeed, the use of AI has raised its share of concerns. Careers once considered safe from automation are now at risk of replacement. AI art, for example, is created by training on assets created by others — something that critics deride as creative theft.
AI can also create fake footage of people and events, known as “deep fakes.” And then there is the risk that an AI will exhibit bias based on the data used to train it, or that it will tell falsehoods if it pulls bad information online.
The experts agree: Knowing how to properly use AI is key.
“You need to be very clear and strong with your prompt engineering, which is an entire skill set in itself,” Bliss said. “This helps ensure that the AI is not hallucinating and creating things out of nowhere, which is something we saw recently in Australia, where a researcher flagged multiple AI-generated errors in a government report, including references to studies and research papers that didn’t exist. That’s part of the risk.”
“Ultimately, you should use AI, but you must do so responsibly and ethically,” Bliss said. “There should be significant training, as well as assurance that there are humans integrated throughout the entire process, from the start with the prompt to reviewing it. There must be significant human intervention — that’s the only way the public is going to trust it.”
Madison Heights Mayor Roslyn Grafstein praised the work of ITAC on this effort.
“Mark Bliss started ITAC in his first term on council and for the last decade it has been a big part of keeping Madison Heights moving forward,” Grafstein said. “I appreciate the work they have done to modernize how we use technology and improve cybersecurity.”
Madison Heights City Councilmember Quinn Wright agreed.
“I really appreciate the work ITAC is doing to modernize our city’s approach to technology and AI policy. Their efforts are helping us strengthen cybersecurity, improve accessibility in our public meetings, and build a more efficient, connected Madison Heights,” he said. “This foresight is exactly what we need to keep our city moving into the future.”
