How London is using AI to crack down on short-term rental scofflaws

Short-term rental landlords have paid $85,000 in fines in the last two years, caught largely using an artificial intelligence (AI) tool that scrapes Internet data and compares it to locally-held records. 

London has 268 active, licenced short-term-rental listings, largely on platforms such as AirBnB and Vrbo. Another 28 applications for licences are pending, officials say. The AI program has been used since October 2022, when restrictions on short-term rentals went into effect in the city. 

"The program scans over 60 of the world's largest short-term accommodation websites, extracting relevant information to narrow down addresses and owner matches," said Nicole Musicco, the city manager in charge of the program. 

"If the program recognizes that one owner has multiple properties, for example, then an enforcement letter is generated through the system and sent to the owners telling them they are not permitted to have more than one short-term accommodation in the city. After that, our bylaw enforcement team would follow up with an inspection." 

Since the program launched, 171 fines have been issued, costing an average of just under $500. 

Most rule-breakers flagged by AI

London's rules require a short-term rental to be the homeowner's primary residence and must be licensed, which costs $193 per year. No applications for licences have been refused this year, Musicco said. 

Most of the fines have been generated because of a flag by the AI software, she added. "It's an incredible tool. It would have taken a lot of staff resources to attempt to locate all of the possible short-term accommodations across the city by address, and then connect them to ownership information." 

Busloads of sorority girls, parties frustrate neighbours of London, Ont., Airbnb-rented home

2 months ago

Duration 3:08

One London home has often been rented on Airbnb for weekend parties featuring busloads of sorority girls — and it’s frustrating the relatively quiet Ontario neighbourhood. Neighbours say it's a stark example of how short-term rentals don't work in residential neighbourhoods and create havoc for those living around them.

In the past, bylaw officers would have to rent a short-term accommodation to get the address, and then go inspect it, she added. 

It's common for municipalities that to use third-party AI software to help them track what's going on in the short-term rental market, said Thorben Wieditz, the executive director of FairBnB, which supports short-term rental regulations in Canada. 

"These platforms can see who rents where, how many times a year, the intensity of use, and in places where there are permit systems in place, they can also identify properties that are being rented online without a valid permit number," he said. 

The platforms can't detect landlords who work hard to thwart the laws, either by registering properties in someone else's name or changing drivers' licences to make it look like they live at a particular address, Wieditz said. 

"There are a number of ways to get around the rules," he said. "Generally, we see that short-term rental regulations are fairly effective, but there's certainly a number of folks that are very dedicated to making a lot of money on this market that have figured out ways of getting around existing rules and regulations." 

A recent report has found that short-term rental listings can generate five to six times the amount of the average rent for an apartment in Ontario, he said. "It is literally the case that short-term guests and tenants can out-compete long-term residents on a nightly basis."