Addressing AI use in the Legal Profession

Artificial intelligence (AI) is becoming a routine part of legal practice. Lawyers use generative AI to conduct legal research, draft documents, and perform administrative functions. AI tools can reduce the time required to perform routine tasks which lowers client costs and expands access to justice. However, its use also raises concerns around client confidentiality, data security, and reliability of work products. Despite these risks, the Rules of Professional Conduct (“the Rules”) do not provide lawyers with clear guidance on how to use AI. This regulatory ambiguity forces lawyers into a dilemma–how can they employ AI in their legal practice while following their professional and ethical obligations?  

The Law Society of Ontario (LSO) designed the Rules to protect the public interest and help guide ethical decision-making. Several provisions directly affect AI use. Rule 3.6-1 requires that a lawyer “shall” charge a reasonable fee while Rule 4.1-1 mandates that lawyers “shall” make legal services available in an efficient and convenient way. At the same time, Rule 3.3-1 imposes a strict duty of confidentiality. Together, these provisions require lawyers to deliver efficient and affordable services while safeguarding client information.  

These competing obligations create regulatory tension. Efficiency and reasonable billing operate as enforceable obligations. If AI can reduce the time needed for research, drafting, and document review, a lawyer who declines to adopt these tools may have difficulty justifying higher fees under Rule 3.6-1. At the same time, the duty to retain client confidentiality is a fundamental aspect of the lawyer-client relationship. Many generative AI models process data through third-party infrastructure, store data in foreign jurisdictions, retain prompts and use user inputs to train models. The Rules do not specify what level of vendor due diligence, data anonymization, or contractual protection satisfies a lawyer’s confidentiality obligations. Without clear regulatory standards governing data retention or permissible uses, lawyers must balance efficiency pressures with confidentiality risks on their own.  

The LSO should resolve this ambiguity. In 2022, it amended the commentary to Rule 3.1-2 governing competence to state that lawyers “should” understand the benefits and risks associated with relevant technologies. The amendment recognizes technological literacy as part of professional competence but does not establish standards for AI use. The Law Society needs to articulate minimum due diligence standards for AI use, including requirements governing data residency, data retention, encryption safeguards, and contractual protections. Further, it should distinguish between publicly accessible large language models such as ChatGPT and enterprise systems that offer enforceable contractual assurances.  

Until the LSO provides clear guidance, lawyers should confine the use of AI tools to processes that do not require disclosure of confidential information. General legal research, document summarization, and administrative support functions allow lawyers to reap the benefits of AI without compromising confidentiality. Technological innovation will continue to redefine legal practice. Without concrete guidance, lawyers must carefully balance competing professional obligations without a proper framework to accommodate them. 

The opinion is the author’s, and does not necessarily reflect CIPPIC's policy position.