Canadian courts have not yet considered applying the law of privilege or professional secrecy to inputs by lawyers into generative AI tools and the resulting outputs. The current best practices suggest that lawyers should disclose generative AI use to clients and should not input privileged information into generative AI tools.

The Canadian Bar Association published guidelines relating to AI use. The guidelines urge lawyers to consider disclosing to clients if they intend to use generative AI and to provide explanations about how this technology will be used. Disclosure should include information about any benefits and risks related to AI use, including risks related to loss of privilege.

Recently, the Law Society of Ontario (LSO) published a white paper that provides guidance on the use of generative AI. The LSO suggests the following best practices:

• Lawyers should not input confidential or privileged information into generative AI tools without ensuring that adequate security measures are in place.
• Lawyers should ensure that they understand the privacy and data security settings of the AI tool and its limitations.
• Even where a lawyer anonymizes input data, there may be residual risks, as AI may be able to piece together information from the anonymized facts.

Law societies in other Canadian provinces have also published guidance on AI use, including in Alberta, British Columbia, Manitoba and Saskatchewan.