It’s science fiction no more.
From garbage cans to flood prevention, the Town of Innisfil is finding ways to use artificial intelligence (AI) and data analysis to help make local government run more efficiently. In late summer, chief financial officer Mike Melinyshyn spoke to a packed room at the Association of Municipalities of Ontario conference in Ottawa and stressed the need for towns and cities to adapt to the modern world and embrace AI.
“This is fun stuff for me,” he said during a recent interview with InnisfilToday. “To me, innovation is an organizational culture thing. Innisfil has really embraced it. How can AI be used to better municipal service delivery?”
For Melinyshyn, it’s about developing smarter, faster and more inexpensive services, which ultimately benefit staff and residents.
Here are three projects, in various stages of development, that have been undertaken by the town and use AI technology:
Not your average garbage can
This first example came out of the COVID-19 pandemic, when more activities were moved outdoors but there was a hesitation by workers toward handling waste.
So, the town partnered with Canadian tech startup Intelight Innovations on a pilot program to test garbage can sensors in outdoor receptacles across the municipality.
“Our garbage cans, if you look at them, they look like they’re sitting three or four feet off the ground,” Melinyshyn said. “Those garbage cans actually go 10 feet into the ground. The operations team literally drives to each garbage can, open it and see if it’s empty. If it’s full, they get a specialized truck with a hoist in. It’s pretty time consuming. We partnered to put sensors in to detect the level of garbage. Intelight created an app for us that tracks the data and feeds it back to the operations team. We know when the garbage cans need to be emptied.”
A second phase of the project will focus on route optimization for municipal staff.
“Innisfil’s huge and we have parks scattered all over the place,” Melinyshyn said. “They don’t have to stop at a garbage can that’s half full.”
A flood of ideas
An average rainstorm or something more damaging?
Down the road, Innisfil staff may have a better sense of precisely when and where flooding is about to happen. Sensors initially employed for testing soil moisture in garden beds and improving watering techniques are being adapted for use in flood-prone areas of the town. The plan, still in the development stage and called “aspirational” by Melinyshyn, could eventually see sensors placed in ditches and connected into nearby stormwater pond levels to do “really predictive” modelling.
“If we can sense how much water’s building up in our soil, stormwater ponds and ditches, and tie that into predictive modelling with the weather, we’ll be able to be better prepared for floods,” he said. “Right now, our operations team drives out, looks at the level of ditches and if they’re not high enough, they drive back. They don’t have to drive out anymore if we can get this technology up and running. We’ll be able to tell when these instances are about to happen.”
The town is also partnering with the University of Toronto Urban Data Research Centre to create digital replications of the municipality that will eventually be used to help with climate-action planning and flood management.
Chat GPT — for government
Innisfil is one of six North American municipalities now using the new generative AI tool GovAI.com, which is described as a “quasi-LinkedIn” for public-sector workers.
Here, staff can research, share ideas and build connections with others in similar positions throughout the continent.
“It’ll generate an output based on my position in my organizations,” Melinyshyn said. “It’s also got the functionality that you can do some networking and collaborating. It’s a really powerful tool to help us do our work.”
But unlike other similar tools such as ChatGPT, GovAI.com uses a datashield blocker to ensure it doesn’t retain personal, identifiable information.
“ChatGPT leaves a postcard to the world. Ours doesn’t do that,” Melinyshyn said. “It stops it from entering the model. We’re using these AI tools responsibly.”
Melinyshyn notes that, save for staff time, the municipality has spent “zero dollars” on these aforementioned programs.
“It’s all about collaboration,” he said.
With his sights set squarely on the future, Melinyshyn sees AI eventually being used to develop predictive modelling for road maintenance and repair schedules, create digital representations of infrastructure like parks — allowing residents to virtually walk through and critique before construction starts — and employ voice replication technology to translate council meetings into dozens of languages in real time.
“The key messaging is that we’re trying to do it properly, transparently and protecting personal, identifiable information,” he said. “It is a powerful tool that can be used for good if managed properly. The guardrails need to be in place.”
When asked whether AI will lead to job losses at the town, Melinyshyn referenced the World Economic Forum. It predicted 85 million lost across the globe by 2025, but anticipated 95 million will be created in their place.
“It is no different than electricity replacing steam; it’s just a different tool,” he said. “We have significant pressures on us to keep costs down. That’s not because we’re hiring hundreds of people. We’re a growth community; we need some bodies. But we can get people doing their work smarter and faster with technology to free you up to do things of value. Yes, there will be a shift in skill set. It’s very gradual. But people will ... do better work.”