At the state and local level, adopting AI remains only one of a long list of IT priorities. Even when AI is near the top, data and systems often need updates to prepare for it. Scot Barker, Chief Innovation Officer for Burlington, Vermont, carefully considers each AI application before using it. “I really have to dive into how I implement new technology on a limited budget,” he said. “There’s not a lot of room for me to say ‘It will be cool to explore this.'”
The Journey So Far
In 2023, part of Barker’s mission was to “demystify” AI for Burlington government colleagues. “I talk about the demystification, because it was a matter of setting people’s minds at ease,” he said. “If you’ve used a smart speaker, you’ve been using AI at some level. The ‘Terminator’ robots aren’t around the corner.”
In 2024, the task shifted to setting formal policy. “I don’t think you can do AI without solid policy. How are you using it? How are you restricting your use? How are you annotating [it] when you’ve used it?” Barker asked.
He is also looking for applications where AI could have the most cost-effective impact. “We’re exploring ways we can leverage robotic process automation,” he said. For example, an automation could enter invoices into the accounting system rather than people typing them in by hand.
Beyond AI, Burlington is rolling out a new website that uses platform-as-a-service technology. Although that system will improve cybersecurity and accessibility and allow the site to adapt better to the device it’s viewed on, there are trade-offs, Barker said. “The new site will definitely modernize the front end in terms of look and feel, but we’re going to lose some functionality, some customization capability. It’s change management.”
On the Horizon
Barker is gaming out ways to integrate about eight terabytes of public records into an AI-driven chatbot that could answer sophisticated questions such as “Can you give me all the documentation for the Great Streets/ Main Streets project?”
Because digital records often contain multiple copies of documents, including outdated versions, Barker estimates that only about half that data is unique. Deduplication — winnowing out the extras without losing needed records — is a project that AI renders far more feasible. “And then we’ve got paper documents that go back forever,” he added. “How do I digitize them? How do I make them part of the overall data repository?”
Barker is looking at how AI can help in other areas, too. For instance, the city’s enterprise resource planning (ERP) system is also nearing the end of its useful life, Barker said, and “AI will certainly play a role in what our next ERP looks like.”
Additionally, he’s always aiming to boost cybersecurity. “If you ask your CIO what keeps him or her up at night and the first answer isn’t cybersecurity, you probably have the wrong CIO,” he said. “We’ve got a lot of new cyber initiatives going on over the next six to eight months.”
But Barker knows these issues aren’t unique to Burlington. “I don’t think we’re super outside the norm,” he said. “A lot of people are in the same situation.”
This article appeared in our guide, “Going Places: Priorities for State and Local Tech.” To read more about how local governments are making the most of technology, download the guide.
Leave a Reply
You must be logged in to post a comment.