Courts Continue to Reign in AI Hallucinations

As we have recently discussed, Ontario Courts are quick to reign in situations where AI is used inappropriately -- or even forbid its use altogether. In a recent Direction re: Defence Final Submissions, in R. v. Chand (2025 ONCJ 282), Justice Kenkel ordered counsel for the defendant to personally prepare a new set of defence submissions without the use of AI software for legal research. In this case, Justice Kenkel noted “serious problems” with defence submissions. One of the cases appeared to be fictitious. Other citations lead to unrelated cases, or to cases that did not support the points being made. The errors were “numerous and substantial.”

Here are two takeaways for e-Discovery professionals:

1. Verify - Even if AI generated results look legitimate, always verify the results. Lawyers must review records to verify AI-generated outputs.
2. AI is a Tool, Not a Substitute - Generative AI is a powerful tool but not a replacement for legal judgment.

The legal profession is at a crossroads. As technology transforms how we research and do our work, the responsibility to ensure accuracy and integrity has never been greater.

So next time you use AI, remember: courts are watching, and so is the future of the legal profession.