Image Source: AI-generated using Canva Magic Media
By
Just a few years ago, Texas was ablaze in wildfires. More than 200,000 acres burned in forests throughout the state. City leaders knew something drastic had to be done and were open to new ideas.
Public power coordinators brainstormed solutions and turned their focus on new technology. Most of it included artificial intelligence.
Could AI solve this serious problem — and other situations where governments needed to act faster and more effectively than most human resources could muster? The answer, for Austin and a growing number of cities and towns throughout the United States, is yes.
AI is the emerging technology that everyone, both in and outside of government, is talking about. AI-powered machines and algorithms are programmed with a human intelligence simulation that helps them learn and refine their outcomes over time. It’s exciting to some and terrifying to others.
For Chris Vetromile, the new Wildfire Mitigation Manager for the publicly run Austin Energy, AI has been a lifesaver — literally.
“It’s the wave of the future in the sense that it will help with everyday tasks and things we haven’t even thought about before,” he said. “There’s the scary AI, the computer becoming human like the Will Smith movie, but this isn’t about a robot taking over. This is a computer algorithm, a really smart computer processing data.”
Austin’s new system, which has only been fully online since August 2024, consists of 13 stations and 26 high-definition cameras that use AI to detect the first sign of smoke, triangulate its location precisely, and alert emergency managers within seconds. The system, a turnkey operation installed by Pano AI, covers 437 square miles in a way that humans never could.
“The technology looks at the camera view and detects anything that looks like smoke. AI processes it and determines if it’s smoke, fog, smog, or the sun reflecting off the building,” Mr. Vetromile explained. “If it’s smoke, the alert gets routed to a call center with a live human, and they hit the button to give us the alert.”
During most of 2024, Mr. Vetromile said they’ve been training the algorithm to know exactly what smoke looks like. Using integrated weather data, GPS positioning, and controlled burns, officials tested the system’s ability to locate the origin of the blaze.
He said that when there’s a small burn on private land, the AI-fueled system usually beats any call to 911. “That saves time and resources that may have been spent investigating the report, and it’s certainly better than stationing a staff member to watch cameras or sit in fire towers 24 hours a day.”
“It’s doing its job and doing it well,” Mr. Vetromile said. “We can get the firefighters there faster, too, because AI can confirm the location. When they can put the fire out fast, it won’t have the destruction it’s capable of.”
Image Source: Pano AI
The new system came up at an Austin Council meeting just a couple of months after the mayor declared a state of emergency due to the rampant wildfires. Mr. Vetromile said there was no pushback with the joint proposal by Austin Energy, the Austin Fire Department, and the surrounding county Emergency Service District Fire Departments.
The system cost $1.175 million for two years, which the council approved in October 2023 as a new line item within the public utility’s annual budget.
“On Labor Day 2011, there were nine fires within Travis County. It was the most destructive wildfire in U.S. history in the amount of housing values burnt,” he said. “City Council was well aware of that and had been trying to figure out mitigation efforts. Pano AI fit in perfectly.”
Preventing fires is not the only way AI is helping local governments. David Solaro, the assistant county manager of Washoe County in Reno, Nevada, recently unveiled a generative AI tool that drastically reduced the staff time needed for searching records and crafting reports for the County’s elected officials. They named it Madison AI after the fourth President of the United States.
The system is an internal search engine that functions similarly to ChatGPT as a knowledge assistant. It pulls up government documents and provides references to meeting minutes to double-check results. So unlike ChatGPT, there are no hallucinations or false results.
The system cost $150,000 to create from scratch by On Strategy, a Reno-based tech firm. With the funding coming from the general fund, it wasn’t enough to require specific board action for approval. Still, Mr. Solaro said he sat down with each commissioner to explain the plan.
He had to do the same with his staff, who expressed many common fears surrounding AI.
“Every employee saw on the news how AI will take their job. When I showed it to them, they instantly understood,” he said. “This isn’t about taking jobs. It’s about helping you do your job faster and easier.”
Madison AI wasn’t the first way that Washoe County attempted to integrate machine learning into their workflow, Mr. Solaro said.
Initially, they had hoped that a generative AI tool could help staff reduce the time they spent processing short-term rental permits. However, as they reviewed the process, they discovered that government workers spent most of their time helping citizens who had questions while filling out their applications.
“We couldn’t shave time with generative AI there, but now we are working with the permitting software to make it easier and intuitive. Once we get that piece taken care of, generative AI can step in to review the checklist,” he said.
He said that other municipalities and governmental agencies in and around Washoe County, such as the city of Reno, Clark County, Santa Clara, and the local regional transportation authority, have reached out to On Strategy to develop a similar internal system. Since they’ll pay Washoe County for the product license, Mr. Solaro said they expect to earn back their initial investment by the end of the first quarter of 2025.
Those additional funds could help other communities expand into AI as well.
“We may help smaller governments in the state to utilize this by using the income to pay for that,” he said. “We got into AI because we wanted to make sure our government could provide for their citizens more easily. It’s all about good governance.”
Mr. Solaro said he encouraged any government officials interested in pursuing AI to create policies and ensure that the use of artificial intelligence fits into the strategic plan.
For example, their AI policy requires staff to fact-check Madison AI results. They should also never input information that isn’t yet available to the public in the system. He added that the focus remains on data security at all times.
What’s next in AI? Mr. Solaro is hopeful he can find an AI tool that helps him sort the hundreds of emails he receives every day. They’ve allocated money in next year’s budget to experiment and find out.
“We’re cautiously optimistic there are more good tools that they do what they say and are helpful for us,” he said. “My biggest advice is to not be afraid to try it. Your employees are probably using it, and there’s a ton of good that can be done with it.”
Magazine
Playmaker Events
Connect with playmaker
