More than a third of Canadian rental and property listings are now generated with artificial intelligence, according to a May 2026 report by AI detection firm Originality.ai. The shift changes how properties are marketed and introduces new risks of inaccurate descriptions that could mislead buyers.
AI adoption in property listings
The study examined text from more than 15,000 rentals and 56,000 sales listings in the first two weeks of May 2026. Researchers found that 37 percent of the listings likely used AI, while 30 percent appeared human-written. The remaining 33 percent were inconclusive because they fell below the 100-word minimum required for accurate detection.
Property sales ads saw higher adoption rates, with 41 percent flagged as computer-generated, compared to just 21 percent of rental ads. Agents looking to integrate these tools can explore the AI Learning Path for Real Estate Brokers to understand best practices.
Time savings and accuracy risks
Real estate agents say the technology significantly reduces the time spent writing descriptions. Calgary-based realtor Amanda Ku said using AI to complete listings saves her up to 30 minutes in a process that usually takes between three and five hours. The tools also help her identify neighborhood amenities like schools and trails.
Michael Fraiman, a researcher on the Originality.ai study, said the technology fits the industry well. "I haven't worked in real estate, but I imagine the job is very social, very personable and interactive sales-based, and not so much a writing job," Fraiman said. "So it makes perfect sense and it's a defensible use case for a realtor to turn to AI."
However, Fraiman warned that AI can generate incorrect information. "There are examples on Reddit of people who found listings that were straight-up incorrect where AI inserted hallucinations," he said. He added that agents must carefully edit the output and check it against reality.
Identifying machine-generated text and regional trends
Detection tools look for specific linguistic patterns to separate human writing from machine-generated text. Informal abbreviations like "br" for bedroom strongly indicate a human author, appearing in AI-flagged ads only nine percent of the time. Multiple consecutive exclamation marks also point to human writers, whereas AI tends to use a single exclamation mark or none at all.
Vocabulary choices provide additional clues. Words like "exceptional" and "convenience" appeared in computer-written ads two-thirds of the time. Conversely, terms such as "must-see," "super," and "gorgeous" appeared in human-written listings, showing up in AI-flagged ads just 19 percent of the time. Renee Sieber, an associate professor of geography at McGill University studying AI, said em-dashes have become a classic giveaway of machine-generated text, while multiple exclamation marks imply human writing.
AI adoption varies significantly by city. Calgary recorded the highest proportion of AI-generated ads at 70 percent. Montreal had the lowest at seven percent, while Vancouver sat at 31 percent. The low detection rate in Montreal stems partly from language barriers in detection tools and the brevity of Quebec real estate descriptions. Professionals exploring regional AI applications can review broader trends on the AI for Real Estate & Construction tag page.
Why this matters for real estate and construction professionals
The data shows that AI is already drafting a significant portion of property descriptions, but it cannot replace human oversight. Agents who rely on these tools must treat the output as a first draft rather than a final product. Verifying every factual claim, especially regarding neighborhood amenities and property dimensions, protects both the buyer and the agent's professional reputation.
Your membership also unlocks:
