Vancouver Property Owner Faces Scrutiny Over AI-Driven Tax Appeal

The owner of a valued property in Vancouver has successfully reduced his tax assessment by nearly $1 million, but his methods have drawn scrutiny from the Property Assessment Appeal Board of British Columbia. Fu D. Ren, owner of an 18,312-square-foot lot in the Mount Pleasant neighbourhood, contested an assessment of $19,082,000 and argued it should be lowered to an estimated market value of $10 million. The board found that Ren may have breached its code of conduct, potentially by using artificial intelligence to support his claim.

In a recent decision, the board’s chair, John Bridal, highlighted that Ren’s submission included legal citations that could not be verified. “The appellant’s submission includes quotations from legal case citations that do not exist,” Bridal stated. He further noted that these discrepancies might indicate either carelessness in the use of AI—leading to what are termed “hallucinations”—or intentional falsification of information.

The controversy arose during Ren’s appeal against the 2025 assessment of his property, which had previously suffered damage from a series of fires. By August 9, 2024, the city mandated the demolition of the remaining structure, which was completed by November 20, 2024. Ren contested the assessment, claiming that the demolition and cleanup costs, totalling $2,030,287.14, should be factored into the property’s value.

In his appeal, Ren cited three decisions from the board and B.C. Courts to support a downward adjustment based on market resistance. However, the board found that these references were also unverifiable. The decision states, “The assessor points out that none of the cited cases can be found on the Canadian Legal Information Institute or board websites.”

Ren’s methodology involved applying a 20 percent discount to a suggested list price of $12 million, arriving at a calculated market value of $9.6 million. However, Bridal concluded that Ren had not presented a reasonable basis to support this valuation, branding the supporting case citations as incorrect and potentially falsified.

The board emphasized the importance of transparency when utilizing AI in legal submissions. Bridal noted that Ren failed to disclose any reliance on artificial intelligence, which is mandated by the board’s code. He stated, “Whether the submission in question was accidental from the poor use of artificial intelligence or a purposeful falsification, doesn’t make a difference.”

The board ultimately ordered that Ren’s property assessment be adjusted down to $18,144,000. Bridal indicated that further submissions regarding potential cost awards against Ren would be considered, reflecting the additional research and investigation necessitated by the misleading submission.

Legal experts have weighed in on the implications of this case. Wayne MacKay, a professor emeritus at Dalhousie University’s Schulich School of Law, emphasized that individuals using AI for legal arguments must ensure the accuracy of their references. “You can use it at your peril in the sense that you’re ultimately responsible for anything it says, both the content and its sourcing,” MacKay commented.

As the use of artificial intelligence becomes more prevalent in various fields, including real estate and law, this case serves as a cautionary tale about the potential pitfalls of relying on such technology without due diligence.