Quebec Bar urges caution when using artificial intelligence

The Montreal courthouse on Thursday, Sept. 19, 2024. (Graham Hughes/The Canadian Press)

Machines will never replace people, says the Quebec Bar. The bar association is urging caution when it comes to the use of artificial intelligence.

Catherine Claveau, president of the Quebec Bar, is warning the public “not to take shortcuts that might be tempting by using artificial intelligence (AI) tools, which cannot replace genuine assistance from a legal professional.”

“The main mission of the Quebec Bar is to protect the public. There are ethical and deontological issues surrounding the legal profession that are not covered by a machine. You have to be wary. It can never be pure legal opinion,” she says in an interview with The Canadian Press.

The bar has launched a national advertising campaign to make people aware of the risks.

“There are tools available to everyone that give you the impression that you don't need to go any further because they give you an answer to your question. The message we want to get across is that the answer you get is not entirely adapted to your context, because the data is depersonalised. There are nuances that cannot be provided by software,” says Claveau.

She points out that “the stakes are high, because this can lead to a loss of rights.”

There are also potential problems of confidentiality, she points out.

“If you do business with a lawyer, all your confidential personal information will be kept. If you give confidential personal information to the machine, you don't know how far it can go. There is a risk of it going astray and not being protected. So you have to be careful before using this kind of software,” Claveau says.

‘We encourage it, we don't forbid it’.

AI tools can be an interesting starting point for the uninitiated.

“The number one piece of advice is to say: if you have a starting point, go with that and call a lawyer to continue the analysis with you. Don't just rely on that because you could end up in a complicated situation.”

Claveau points out that, unlike AI tools, lawyers are accountable.

“If, for example, a lawyer makes a mistake, there is professional liability insurance so that, if it causes you trouble or damages, you will be able to get compensation, which is not the case if the loss of rights comes from a machine.”

Despite these reservations, the bar is not demonizing AI. The association has ruled that the technology can prove useful.

“When we started talking a few years ago about the arrival of artificial intelligence, we were proactive and asked ourselves whether we should be for or against it,” says Claveau. “We quickly realised that we couldn't be against it. Artificial intelligence is now a part of our practices. We encourage it, we don't forbid it.”

To provide a framework for legal professionals, the bar created a guide to best practice in this area last November. The document will be updated in line with technological advances.

“Obviously, we'll continue to keep abreast of developments because they're evolving fast. We'll be improving the guide over time to provide a proper framework for this phenomenon,” says Claveau.

The bar has also hired a lawyer specialising in this area to help “develop both our position as a professional order and our orientation.”

Not a panacea

Many of the bar’s 30,500 or so members use AI to save time.

Claveau cites documentary research — which often requires lawyers to sift through a large volume of doctrine, i.e. all the works that set out and interpret the law — as an example.

“We often rely on precedents in case law. There are some very serious companies, like CanLII, that are developing increasingly sophisticated research software that makes increasing use of artificial intelligence,” she says. “Cross-referencing certain data can really save us a lot of time rather than spending hours in the library.”

She also cites court preparation: “When we have to prepare examinations, cross-examinations or a pleading plan, we can give data to the machine, which will help us to structure our thinking. In terms of form, we can save a lot of time.”

But Claveau says this also requires caution.

“A human must always go back over the machine to see if the proposed model really fits the customer's situation. You have to continue to analyse, consult your colleagues, work as a team and validate yourself,” she says.

Claveau insists that “it's not a panacea” and “they are tools to make our work easier and lighter, but we have to remain vigilant and use them sparingly.”

This report by The Canadian Press was first published in French on Jan. 12, 2025.