Law Society AI Guidance: What Canadian Lawyers Need to Know by Province

Core duties that apply across Canada

The Canadian national baseline is the FLSC’s Model Code of Professional Conduct. While the document does not specifically regulate AI, it sets out the professional duties that apply when lawyers use it. The three most relevant are Rule 3.1 (Competence), Rule 3.3 (Confidentiality), and Rule 6.1 (Supervision).

While each province’s law society has taken its own approach to generative AI, the guidance is largely consistent. Regulators generally treat AI as another tool that lawyers must use in accordance with their existing professional obligations, rather than creating entirely new AI-specific rulebooks. Across jurisdictions, several common themes emerge:

• Competence. Lawyers should understand the capabilities and limitations of the AI tools they use.
• Confidentiality. Client and privileged information must be protected.
• Supervision. Lawyers remain responsible for reviewing and verifying AI-generated work.
• Transparency. In some circumstances, lawyers may need to disclose or explain their use of AI.

These shared principles set the baseline for AI use in legal practice across Canada. What that looks like in practice varies by jurisdiction. 

Provincial law societies have supplemented the Model Code with their own AI-specific guidance.

What do Canadian law societies say about AI? 

In general, law societies are taking a measured approach to AI. They are not banning lawyers from using it. They are, however, making it clear that lawyers remain responsible for meeting their professional obligations when using these tools.

Province-by-province breakdown of law society AI guidance

Practising in Ontario, British Columbia, Alberta, or elsewhere? Below, we summarize published guidance across key Canadian jurisdictions and highlight what’s notable in each.

Province/territory	Guidance	Key takeaways
British Columbia	Guidance on Professional Responsibility and Generative AI	If AI completes work in less time, time savings must be reflected in fair and reasonable fees. Confidentiality concerns may require shifting from public to secure or proprietary AI tools.
Ontario	Licensee use of generative artificial intelligence	Client-facing AI tools risk creating unintended lawyer-client relationships. They should be avoided unless appropriate controls are in place. Disclosure of AI use is not generally required, but may be necessary depending on client expectations, court rules, or data sensitivity.
Alberta	The Generative AI Playbook	Establish firm-wide policies around AI; assume junior lawyers and staff already use it.  Alberta Courts require AI-generated citations to be verified against authoritative sources like CanLII; the Federal Court requires written disclosure in the opening paragraph of every AI-assisted filing.
Saskatchewan	Generative AI Guidelines	Strong emphasis on continuous learning: AI competence is an ongoing obligation, not a one-time exercise.
Manitoba	Generative AI Guidelines	AI is a tool to assist legal work, not a substitute for professional judgment. Responsibility for legal advice, strategy, and work product always remains with the lawyer.
Northwest Territories	Guidelines for the Use of Generative AI in the Practice of Law	AI tools may cite American law over Canadian law; they must be carefully checked for jurisdictional accuracy and bias. Fee reductions should reflect AI-driven efficiency, and AI-related costs may need to be transparently disclosed to clients.
Nova Scotia	Artificial Intelligence in the Practice of Law – What is AI and can I or should I use it in my practice?	A plain-language primer aimed at lawyers less familiar with AI, covering what the technology is and whether and how to use it responsibly. The core duties of competence, confidentiality, and supervision apply as they would to any other tool.
Quebec	Generative Artificial Intelligence – Practical Guide for Responsible Use	Pairs a practical guide with supporting tools, with a strong focus on transparency and privacy. Quebec’s Law 25 adds transparency obligations where AI factors into automated decision-making or the handling of personal information.
Prince Edward Island	Law Society of PEI’s Artificial Intelligence Guidelines	Covers competence, confidentiality, and supervision, and goes further on billing. Lawyers are cautioned against “fee windfalls” that arise when AI-driven efficiencies are passed to clients as billed time.

British Columbia

The Law Society of British Columbia (LSBC) has issued some of Canada’s most comprehensive guidance on generative AI through its Guidance on Professional Responsibility and Generative AI, making clear that AI does not alter a lawyer’s existing professional obligations. Duties relating to competence, confidentiality, honesty and candour, and supervision apply as they would with any other technology.

Lawyers must understand the capabilities and limitations of AI tools, independently verify any AI-generated output, and know how that system collects, stores, and uses data before entering client information into it. In some circumstances, informed client consent may be required before confidential information is shared with an AI system. For example, this could arise if a lawyer wishes to upload confidential client information to a generative AI platform that retains user inputs or uses them to train future models.

Ontario

In its 2024 white paper on generative AI, the Law Society of Ontario (LSO) encourages lawyers and paralegals to explore and experiment with AI tools while remaining mindful of their professional obligations. Rather than introducing AI-specific rules, the LSO confirms that existing duties continue to apply.

As of April 2026, lawyers are not required to disclose their use of AI to clients solely because AI was involved in their matter. Guidance continues to evolve, and lawyers should monitor updates through the LSO’s Technology Resource Centre.

Alberta

The Law Society of Alberta’s Generative AI Playbook outlines the benefits, risks, and practical considerations associated with using AI in a law firm setting. Its Code of Conduct explicitly requires lawyers to “develop an understanding of, and ability to use, technology relevant to the nature and area of the lawyer’s practice and responsibilities,” linking AI competence directly to the broader duty of technological competence and positioning it as an indispensable part of competent legal practice.

Saskatchewan

The Law Society of Saskatchewan’s Generative AI Guidelines focus on continuous learning, client transparency, and careful oversight of AI tools. Saskatchewan’s central message is that AI competence is an ongoing obligation: Lawyers must stay informed about new developments, risks, and best practices, rather than treating it as a one-time exercise. 

Manitoba

The Law Society of Manitoba’s Generative AI Guidelines address competence, confidentiality, and supervision, acknowledging the benefits of AI while reinforcing that lawyers remain responsible for complying with their professional and ethical obligations. Lawyers must understand AI’s capabilities and limitations, safeguard confidential and privileged client information, and review and verify any AI-generated content before relying on it. Manitoba’s guidance is notable for its emphasis on accountability.

Northwest Territories

The Law Society of the Northwest Territories’ Guidelines for the Use of Generative AI in the Practice of Law, issued in January 2025, take a pragmatic, principles-based approach, applying existing professional duties directly to AI use rather than creating a separate regulatory regime. Lawyers must understand AI’s capabilities and limitations, protect client confidentiality, verify AI-generated outputs, and remain fully responsible for their work product.

Other provinces with guidance

The Barreau du Québec has published dedicated generative AI resources covering practical guidance, tools, and considerations around transparency and privacy. Nova Scotia has released an AI guide for legal practices addressing the profession’s core obligations in the context of AI use. Prince Edward Island’s AI guidelines cover competence, confidentiality, and supervision. They also address the obligation to avoid “fee windfalls” arising from AI-driven efficiencies.

As AI guidance across Canada continues to evolve, lawyers in jurisdictions not covered above should monitor their law society’s website for updates and consult the FLSC’s Interactive Model Code as a baseline reference.

Key risks lawyers must manage when using AI

AI offers significant benefits for lawyers across all practice sizes, especially for solo practitioners and small- to mid-sized firms. But it also introduces risks that must be actively managed:

• Hallucinations and inaccurate citations. Generative AI tools, especially those not grounded in the law, can fabricate case law, statutes, and quotations, including instances that have led to real consequences in court proceedings.
• Client confidentiality. Entering sensitive or privileged client information into public AI systems can risk unauthorized disclosure. Foreign data access laws, such as the U.S. CLOUD Act (Clarifying Lawful Overseas Use of Data Act), may also affect how data stored on U.S.-based platforms can be accessed. 
• Bias in output. AI systems trained on U.S. or global datasets may produce results that favour American law or reasoning over relevant Canadian authorities.
• Court disclosure requirements. Both the Federal Court of Canada and many provincial courts require disclosure where AI is used in preparing court materials under various practice directions. There is currently no equivalent national rule across other Canadian courts.

Managing these risks is at the heart of the AI guidance issued by Canadian law societies.

Evaluating AI use cases by risk level

Not every AI use carries the same level of risk. Lawyers should also assess whether a given AI use case presents a low, medium, or high level of risk and adjust their approach accordingly.

Risk level	What it involves	Examples
Low	Administrative or preliminary tasks with no confidential client information or legal judgment.	Brainstorming ideas, creating administrative checklists, summarizing publicly available information.
Medium	Client matters that stop short of providing legal advice.	Drafting client emails, creating internal research plans, summarizing legal documents.
High	Legal judgment, client rights, or privileged information.	Legal research, drafting court filings, generating legal advice, reviewing privileged documents, deploying client-facing chatbots.

As the risk level increases, so does the need for lawyer supervision. Regardless of the use case, Canadian law societies consistently emphasize that lawyers are responsible for the work produced and must review and verify AI-generated content before relying on it.

How to verify AI-generated legal work

AI-generated content should never be taken at face value, but what verification looks like depends on the task. A brief administrative checklist may require only a quick scan. By contrast, AI-assisted legal research, court filings, and client communications require careful verification before being relied on, filed, or shared.

When reviewing AI-generated work product, apply the following checklist where applicable:

• Verify every case citation.
• Confirm that cited authorities support the corresponding legal proposition.
• Check that cases and statutes are good law and still current.
• Confirm that the jurisdiction is correct and relevant.
• Compare factual statements against the client file and source documents.
• Review outputs for confidentiality, privilege, and potential bias.
• Ensure that a lawyer reviews and approves the final work product before it is filed, submitted, or sent to a client.

Lawyers should maintain appropriate oversight and exercise independent professional judgment at all times.

How to choose an AI tool that meets your law society’s standards

Beyond understanding law society guidance on AI, the practical question becomes how to choose an AI tool that aligns with your professional obligations.

Most law societies emphasize that competence includes knowing how AI tools handle confidential client information. The LSBC, for example, advises lawyers to understand how an AI system collects, stores, processes, and shares data before entering client information into it.

This is where the distinction between general-purpose AI tools like ChatGPT or Gemini and purpose-built legal AI platforms becomes relevant. While both can be useful, they differ significantly in how they handle data, security, and compliance. When evaluating your options, ask:

• How does the system collect, store, and process client data?
• Is client data used to train the provider’s model?
• Where is data stored, and what privacy rules apply?
• What security controls protect confidential information?
• Does the platform provide audit trails or records of AI-assisted work?
• Can lawyers control who has access to client data and outputs?

These questions are essential because lawyers are responsible for protecting client confidentiality regardless of the technology they use.

Purpose-built legal platforms are often designed with these obligations in mind. They offer features including legal-specific privacy controls, contractual commitments regarding data use, secure document environments, audit logs, and administrative controls that help firms demonstrate compliance. By contrast, lawyers using general-purpose AI tools may need to conduct additional due diligence to determine how client data is handled and what safeguards are necessary to meet their professional obligations.

For example, Manage AI is a purpose-built legal AI assistant that operates within Clio Manage, where client data is never used to train public models, and access is governed by your firm’s existing permissions.

Your obligations to clients

Choosing the right AI tool is only part of the picture. How lawyers communicate with clients about AI use, and how they bill for it, carries its own professional obligations.

Do you need to tell your clients you’re using AI?

Lawyers are not generally required to disclose AI use to clients, but disclosure requirements vary by province. Across jurisdictions, the emerging best practice is to address AI use directly in the retainer agreement rather than handling it on a case-by-case basis. This typically involves a high-level explanation of whether and how AI tools may be used, along with their potential benefits (such as efficiency) and risks, particularly around confidentiality, accuracy, and oversight.

In Quebec, Law 25 (formerly Bill 64) adds transparency obligations where AI may be involved in automated decision-making or the handling of personal information, requiring more structured disclosure in those contexts.

How should lawyers bill for AI-assisted work?

Lawyers should not bill clients for time saved through AI automation. Billing should reflect the lawyer’s actual professional work, including legal analysis, judgment, review, editing, and strategy.

Where AI tools are used, firms should also be transparent about any related software or service costs passed on to clients. The right approach may vary between hourly and fixed-fee arrangements, but the guiding principle is the same: Clients are paying for legal expertise, not automation.

From AI guidance to everyday practice

With 95% of Canadian law firms already using AI in some capacity, the question is no longer whether to adopt these tools. It’s how to do so responsibly. Law societies have given Canadian lawyers the green light to use AI, but that permission comes with clear conditions. Across jurisdictions, guidance consistently relies on four core pillars: competence, confidentiality, supervision, and transparency. These are the same principles that have always governed legal practice in Canada.

Lawyers shouldn’t wait for their province’s AI guidance to be finalized before moving forward. The FLSC Model Code and existing professional obligations already provide a solid framework for using AI responsibly.

Firms that use AI well do so by pairing new tools with the same professional judgment that has always defined competent legal practice. Understanding the rules is the first step. Choosing tools that help you comply with those rules comes next and is equally important. As AI adoption accelerates across the legal profession, the tools you choose should support and strengthen your ethical and professional responsibilities. 

The pace of technological change is unlikely to slow down, but the principles that govern legal services are constant. With the right knowledge, judgment, and tools, you can embrace AI with clarity and confidence.

For a deeper look at the Canadian regulatory landscape, read our AI legal compliance guide. And when you’re ready to put your knowledge into practice, Manage AI can help.

What does my law society say about AI?

Most Canadian law societies permit lawyers to use AI tools, provided they understand how the technology works, protect client information, verify outputs, and comply with their professional obligations. Lawyers remain responsible for all work product, regardless of whether AI assisted in its creation.

Which Canadian law societies have published AI guidance?

Several Canadian law societies have issued guidance on the use of AI, including those in Ontario, British Columbia, Alberta, Saskatchewan, Manitoba, Nova Scotia, PEI, and Quebec. While the details vary by jurisdiction, the guidance generally focuses on four core themes: competence, confidentiality, supervision, and transparency.

Is there a national standard for AI use by Canadian lawyers?

There is no single national standard governing AI use by Canadian lawyers. Regulation remains with provincial and territorial law societies, each of which issues its own guidance. They are all, however, grounded in the shared ethical duties set out in the Federation of Law Societies of Canada (FLSC) Model Code of Professional Conduct.

What are the main risks of using AI in my law practice?

The main risks include inaccurate or fabricated information (known as “hallucinations”), breaches of client confidentiality, privacy concerns, biased outputs, copyright and intellectual property issues, and overreliance on AI-generated content. Because AI tools can produce convincing but incorrect results, lawyers should independently verify all outputs before relying on them.

Do I need to tell my clients I am using AI?

There is generally no blanket rule in Canada requiring lawyers to disclose AI use to clients. However, disclosure may be appropriate where AI use raises issues related to confidentiality, informed consent, or the delivery of legal services.

Can I use ChatGPT or other general AI tools for legal work?

Yes, but with caution. Before entering any client information into a general-purpose AI tool, lawyers should assess the risks, including confidentiality, privacy, accuracy, and data-handling, and confirm the tool can be used in compliance with their professional obligations.