Artificial intelligence for cities and counties

It appears that artificial intelligence (AI) is everything, everywhere and in every product the vendor community would have us buy. However, despite the hype, few dismiss it as just another high-tech fad. Indeed, some believe it is the third (or fourth) wave, depending on how one is counting in the ever-evolving digital age.

As early as 2016, the National Science and Technology Council laid out a strategic plan for federally funded AI. In 2019 the American AI Initiative was promulgated through an executive order setting a strategy for AI development in the U.S., emphasizing American leadership in AI, setting AI research and development priorities, and developing AI regulatory guidelines. And in 2020, the National AI Initiative Act was passed that was aimed to bolster AI research, standards, and education. Today many federal agencies are grappling with AI in many ways and one can see many of these initiatives by simply going to www.aI.gov. Like many innovations, what happens at the federal level impacts and influences state and local government.

With so much conversation about artificial intelligence as the next “big thing” in local government IT, the Public Technology Institute wanted to gain a better understanding on how city and county IT executives view AI. In August, PTI conducted the very first survey on AI directly focused on cities and counties across the United States. Perhaps not surprisingly, this short survey received the highest response rate of any survey in PTI’s 50+ year history. Of those responding, 58 percent of local government IT executives believe AI will bring dramatic change to local government operations and service delivery over the next three years, while 62 percent said they are not currently involved in any projects. A surprising 38 percent said they were.

When it comes to the concerns and benefits of AI, survey participants were asked to rank order from a pre-determined list of seven key areas, from most concerning to least concerning. Not surprising, ethics was ranked No. 1, followed by unauthorized use of personally identifiable information, wrong and harmful decision-making, misinformation/disinformation and bias.

In wanting to learn more about how IT executives feel regarding local government functions and operations and how local governments could benefit most from AI, survey participants were asked to rank each function from most to the least among a list of 10 choices. Based on the average weighting of the data collected, of the top five benefits that were selected, cybersecurity management ranked No. 1. Data analysis was second on the list, followed by citizen engagement, predictive analytics in crime prevention and improved decision making.

The survey went on to ask about AI training for themselves and their staff. Here, 85 percent of IT executives stated that they believe they need training to better understand the implications of AI. It was clear from the study that the overwhelming majority felt that they had much to learn about AI, what to look for and or avoid when it comes to choosing new services or service providers that rely on AI for starters.

Finally, the survey participants were asked about whether their local government has an AI policy in place. Only 4 percent answered yes, with 42 percent responding no. But 33 percent indicated they are working on an AI policy now. It should also be noted that 17 percent responded “We should be developing one.”

Today there are approximately a dozen cities and counties that have developed either AI policies or guidelines. But there is every indication this number will triple by year’s end. Many local government IT executives along with senior public managers have voiced concern regarding current usage of generative AI such as ChatGPT. They have expressed concern over AI use in job hiring/filtering as well as promotions where bias could violate state and federal laws and regulations. Others have expressed concern that government employees may unwittingly place sensitive information into AI systems without appropriate safeguarding of personally identifiable information. One might imagine a public safety unit inputting crime information into an AI system to look for anomalies or patterns that could help reduce or better predict crime. The same could be said of citizen health data with the goal of tracing or tracking outbreaks. Whatever data is entered into an AI system what safeguards defined by policies are in place to protect privacy and confidentiality? Additionally, some have expressed concern over the seeming lack of copyright controls which could lead to lawsuits or misinformation.

Clearly, the explosion of AI interest and current applications outpaces local government efforts to craft sound governance and policies. This leaves government officials trying to keep up.

Generative AI has deservedly become the most concerning AI application given its demonstrable powers to provide (mostly right) answers, perfectly formatted paragraphs, and sentences, and with perfect grammar. There is no doubt AI will grow and learn in both its power of information and new and exciting applications. Good data serves as the active fertilizer for AI to prosper all the while learning new things through every inquiry, response, and new data sources it can tap. When asked what AI can do for local government, one can merely ask ChatGPT itself for a rather contemplative answer for starters.

Editor’s Note: Dr. Shark has previously addressed AI on this site. Here are links to those stories.

How can local governments regulate generative AI—Just ask ChatGPT
https://www.americancityandcounty.com/2023/07/19/how-can-local-governments-regulate-generative-ai-just-ask-chatgpt/

Generative AI expands the race for search and meaning—Implications for society and government
https://www.americancityandcounty.com/2023/02/15/generative-ai-expands-the-race-for-search-and-meaning-implications-for-society-and-government/

Dr. Alan R. Shark is the executive director for the Public Technology Institute (PTI), a division of Fusion Learning Partners; and associate professor for the Schar School of Policy and Government, George Mason University, where in addition is an affiliate faculty member at the Center for Advancing Human-Machine Partnership (CAHMP). Shark is a fellow of the National Academy of Public Administration and co-chair of the Standing Panel on Technology Leadership. A noted author, his most recent textbook, “Technology and Public Management,” was recently published. He is also the host of the popular bi-monthly podcast, Sharkbytes.net.