Burlington pilot project uses AI to aid in building developments

Burlington is using artificial intelligence to determine how to best shape the future of the city. 

The city just wrapped an eight-week pilot project that saw city workers using two AI tools to help design new homes, and aid in hitting housing targets. The first is used to assess zoning bylaws for industrial and commercial buildings, to make sure they have the correct area, parking requirements, and heights. The second is to assess architectural drawings and reference them with the Ontario Building Code. 

“People’s misconception of AI is that it thinks like a human, but it’s really an observation tool,” Chad MacDonald, executive director of digital service and chief information officer for Burlington said. “It has been learning our rules, we’re training it. It will come back to us and tell us that the floor space isn’t within the right limit, and we’ll tell it that it actually is. Then next time it knows better.”

The city used the AI in building reviews for buildings all over town, and the computer program is able to provide detailed breakdowns – room by room – for whether or not the building is up to code. 

Results can be generated in only a few minutes for entire buildings. MacDonald said uploading the plans to the program takes more time than it does for the program to provide results. A human is still needed to ensure regulations are actually followed, and that the AI hasn’t confused itself.

“There’s still some work to be done,we’re still in the early days, but we’ll do an assessment to find whether this is financially viable soon,” MacDonald said. “We’re just experimenting with it.”

The plan is to find ways for the AI to reduce the number of manual exchanges between applicants and city staff, let applicants get instant feedback on proposals, shorten design time, and ensure higher quality submissions, among others. 

MacDonald demonstrated the work that the AI tools can do during last month’s pipeline to permit committee meeting. The standout feature he showcased is the ability to select a lot in Burlington, and tell it that you want a certain type of building – such as a multi-unit home – and having it model hundreds of potential options. The team hopes that this type of AI tool may be used in other ways soon.

“Conservation Halton asked if we could add different Geographic Information System layers, like flood plains, to the technology, and we absolutely can,” MacDonald said. “That will help their assessments, so now how do we help other organizations? We found in some of our experimentation that the significant amount of time spent on getting building permits is due to the quality of the application, and the number of back and forths required. This is meant to cut all of that because we’re providing immediate feedback.”

The city uses a program called MyFiles, an online tool that is designed to speed up the pre-building approval process for applicants, and allows the applicants to view the application in real time as it goes through different stages of review. Before implementing tools such as this, approvals took an average of 17 weeks. They now take an average of 5.5 weeks, with about a third taking less than two weeks to complete. 

It’s not clear when or if these AI tools will be fully implemented in the city’s systems, and make their way past the pilot project phase. MacDonald hopes that the tool can be used across the province to check if building proposals in Ontario are up to code.