1. Home
  2. Resource Centre
  3. Strata 101
  4. Why stratas shouldn’t rely on AI for information

Why stratas shouldn’t rely on AI for information

Artificial intelligence (AI) tools

Artificial Intelligence (AI) has introduced many tools and efficiencies that help us perform tasks at home and at work. Many of us use digital assistants such as Apple Intelligence or Microsoft’s Copliot. Companies are investing billions in Large Language Model products such as Google’s Gemini, Anthropic’s Claude, Meta’s Llama, and xAI’s Grok.

A large language model (LLM) is a type of artificial intelligence. Through machine learning it’s trained to understand and generate human-like text. It learns from billions of words found in books, websites, and other sources to understand how words relate to one another. LLMs such as ChatGPT can be helpful to summarize long documents, draft emails, help with scheduling, create meal plans, plan vacations, assistwith spreadsheet formulas, coding, and more.

AI-powered search engines such as Perplexity, use multiple LLMs to process questions and summarize answers. Perplexity includes the sources, so the user can check the original information and verify its accuracy.

Why does AI produce fake information?

As the American Association for the Advancement of Science explains, a LLM is trained to fake answers it doesn’t know. It makes predictions based on patterns.

Performance benchmarks reward confident guesses and penalize uncertainty. A fake-it-til-you-make-it model ends up ranking better than a model that admits it doesn’t know the answer. Instead, an LLM will confidently spout factually incorrect statements known as hallucinations.

Some AI workers tell their friends and family to stay away from AI. An AI tutor who has worked with Gemini, ChatGPT and Grok said, “We joke that [chatbots] would be great if we could get them to stop lying.”

A study in Australia warns researchers about relying on ChatGPT to speed up their work. It found that 56% of citations were either fabricated or contained errors. Careful checking would reveal a mismatch between what the tool claimed and what the source actually contained. Among fabricated citations, over 60% deceptively linked to actual published papers on completely unrelated topics. The scope of the problem compounds if articles built on phantom sources and false premises are published.

Why do people believe AI?

In a UBC study published in October 2025, researchers found that when giving medical advice, ChatGPT’s human-like tone was empathetic. Users found the AI-generated responses convincing and trustworthy, even when the information wasn’t correct. An AI rater commented, “This is not an ethical robot. It’s just a robot.”

When it comes to accurately diagnosing a health problem, researchers found that services like ChatGPT can be wrong more often than they’re right. Patients increasingly show up to medical appointments, citing diagnoses or treatment suggestions from AI tools. Doctors often spend a lot of time educating them about what’s actually going on.

Even legal professionals are being hoodwinked. The first case I noticed about the misuse of AI tools in court happened right here in BC. In the family law case Zhang v. Chen, 2024 BCSC 285 before the BC Supreme Court in February 2024, lawyer Chong Ke described being mortified and deeply embarrassed that 2 cases she cited did not exist. While Justice Masuhara found Ke was naïve about the risks of using ChatGPT, apologized, and did not intend to deceive the court, he ordered Ke to pay the cost of the additional effort and expenses incurred by the claimant. In future, Ke must advise the court and the opposing parties about submissions generated by AI tools.

In New Jersey, attorney Michael Fourte used hallucinated quotes and citations to defend his client in a case about a loan. When the opposing side filed a request for Fourte to be sanctioned, Fourte’s rebuttal contained another wave of made-up or erroneous citations. Justice Cohen found the denials about using AI didn’t hold water. In addition to sanctions, the matter was submitted to the New Jersey Office of Attorney Ethics.

In June 2025, London’s High Court (UK) gave a stern warning. Lawyers who use AI to cite non-existent cases can be held in contempt of court or even face criminal charges. The decision said that in the most egregious cases, deliberately placing false material before the court amounts to the criminal offence of perverting the course of justice, carrying a maximum sentence of life imprisonment.

Many lawyers are using AI tools responsibly. The BC Law Society provides information and considerations in Guidance on Professional Responsibility and Generative AI. For example, it advises legal professionals to make their client aware of the use of AI, verify all information, and follow any court rules about disclosing when generative AI was used to prepare their submissions.

Don’t use AI tools in a CRT dispute

Here are some cautionary tales for strata councils and owners. While AI tools may seem convincing, their use has backfired for participants in Civil Resolution Tribunal (CRT) cases.

In Geismayr v. The Owners, Strata Plan KAS 1970, 2025 BCCRT 217, the Geismayrs used Microsoft’s Copilot for information. Their submissions referenced 10 legal decisions saying courts ruled that a strata corporation couldn’t force the removal of alterations to a strata lot. Nine of those cases do not exist. One case exists but is not about unauthorized alterations.

Copilot’s analysis of the law was also incorrect. The CRT found that owners cannot reasonably expect retroactive approval for alterations done without the strata’s prior authorization.

In the small claims case Maxwell v. WestJet Airlines Ltd., 2025 BCCRT 1146, the tribunal member noted that the BC Supreme Court and the CRT have discussed the risks of relying on generative AI tools, as the output is often inaccurate. The tribunal member found the information ChatGPT provided Mr. Maxwell was unreliable and gave no weight to his ChatGPT evidence. Find other CRT decisions about AI hallucinations.

The CRT rules prohibit participants from including non-existent cases or legislation in their arguments, such as those created by an AI tool. There could be monetary consequences. Providing false evidence, cases, or legislation created by an AI tool could result in an order to compensate the other party for time spent dealing with a CRT proceeding, and pay their legal expenses. See the CRT web page Can I use AI tools in evidence and arguments?

Protecting personal and strata information

When using ChatGPT, Gemini, Clause and other AI tools, you might upload documents to help write letters, analyze financials, and more.  This raises concerns for strata corporations in regards to the Personal Information Protection Act. Strata councils and owners in possession of documents should take care to safeguard information when using an AI tool. Take a moment to read the AI tool’s privacy policy, terms and conditions. If there are any potential privacy issues, don’t upload the documents. At the very least, learn how to opt out of tools or features that pose privacy concerns.

Advice for strata councils and owners

ChatGPT’s creator says that ChatGPT has never been a substitute for professional legal or medical advice. In October, 2025 the company updated its usage policies to clarify that users of ChatGPT can’t use the service for “tailored advice that requires a license, such as legal or medical advice, without appropriate involvement by a licensed professional.” An AI tool’s responses are based on patterns in data, not on legal reasoning.

Council members are required to “act honestly and in good faith with a view to the best interests of the strata corporation, and exercise the care, diligence and skill of a reasonably prudent person in comparable circumstances.” At council meetings, you make decisions about spending common funds, approving alteration requests, enforcing the bylaws, and much more. Most would agree that it’s not in the best interests of the strata corporation or owners to rely on information or advice from an AI tool.

As the saying goes, you don’t know what you don’t know. Without expertise, you won’t know which parts of the AI-generated answer are accurate and which parts are fake. For example, AI tools have invented parts of the Strata Property Act or Regulation that have never existed. It looks convincing but it can be completely made up. If you use AI tools, check everything. Go to the source. Read it thoroughly. If the tool doesn’t tell you the source, should you rely on the information?

For reliable information, we encourage you to search topics in this Resource Centre. The articles cite relevant parts of the Strata Property Act and direct you to trustworthy websites and resources. And, when you need legal advice, ask a real, human lawyer!

Was this article helpful?

Related Articles

← Resource Centre