Just recently, the Nova Scotia Barristers’ Society (NSBS) released its 42-page AI Guide: Practical considerations when assessing and using AI in legal practice — a text to help lawyers in the province navigate an “evolving landscape,” while continuing to adhere to their professional and ethical obligations.
Right off the hop, the guide recommends lawyers approach AI with both “curiosity and caution.”
“When used thoughtfully, these tools can enhance legal service delivery and access to justice,” it states. “When used carelessly or without oversight, they can mislead courts, breach confidentiality, or perpetuate bias and discrimination.”
The AI guide, it notes, is grounded in the NSBS’s Code of Professional Conduct. It notes AI is increasingly showing up in legal practices as a way of streamlining routine tasks, reducing drafting time and, generally, improving efficiency.
But it warns that AI is “a predictive tool, not an authoritative source,” and that lawyers must remain responsible for verifying the content that AI generates.
Risks of AI, it states, include the fabrication of information (known as “hallucinations”), breaches of confidentiality and the creation of biased or discriminatory output.
The guide also warns of “over-reliance or misuse,” and the “risk of using AI to perform tasks that require legal knowledge, discretion or ethical decision-making.”
“AI should not be treated as a substitute for legal judgment or supervision,” it states.
Later, it cautions yet again on becoming over-reliant on the technology.
“Do not use AI to prepare arguments or legal analysis without meaningful lawyer oversight,” it states. “Submissions are expected to reflect a lawyer’s professional assessment and interpretation of the law.”
Jonathan Saumier, Nova Scotia Barristers’ Society
“We felt it was necessary to separate the hype around AI from what’s real and really couch this whole AI business — just keep it grounded in lawyers’ ethical and professional responsibilities,” said Saumier.
AI is simply “a tool” to help lawyers do their jobs, he said.
“It doesn’t replace you and your professional judgment, and it doesn’t negate or replace the fact that you still need to be aware of your professional and ethical responsibilities when you use these tools.”
Saumier was asked of the guide’s warning about over-reliance, and if there was fear that busy lawyers would end up over-relying on AI, anyway — despite the warnings.
“Absolutely — and I think we’re past the fear,” he said. “I think this fear has materialized. We’ve seen a few cases now — I think a couple in Ontario and one in B.C. — where lawyers were caught for submitting fake case references in briefs that were submitted to court. And that has had serious consequences for those lawyers, whether it was formal reprimands from the court or even, in one case, being asked to pay costs to the other party because of basically wasting the other side’s time, [with them] looking everywhere to find those cases that didn’t exist.”
Back in May, a judge presiding over a divorce case in Ontario’s Superior Court wrapped the knuckles of a lawyer for filing a factum filled with fake, AI-generated case citations. Around the same time, a judge with the province’s Court of Justice reportedly delivered a similar scolding to a lawyer in a criminal case.
And in early 2024, a judge in a B.C. divorce case admonished a lawyer for filing non-existent citations — but in that case the lawyer was ordered to pay costs to the other side for the time it took to discover the citations were fake.
“Over-reliance on AI has happened, and we’ve seen that in those cases,” said Saumier. “And those cases should be taken as stern warnings, and they should reinforce the importance of avoiding over-reliance [and] of keeping a human in the loop at all times.”
If you have any information, story ideas or news tips for Law360 Canada, please contact Terry Davidson at t.davidson@lexisnexis.ca or 905-415-5899.