AI Notetaking in the Legal and Business Context: Does It Risk Confidentiality or Privilege?

Written By Benjamin Reingold, Stephen Burns and Sebastien Gittens

As businesses modernize, many are turning to artificial intelligence (AI) tools to increase productivity and lower costs. One such tool is AI notetaking—automated meeting transcriptions, real-time summaries and searchable records.

These tools have the potential to offer greater efficiency, consistency and integration with enterprise workflows. But for in-house counsel, they also raise legal and ethical considerations—particularly around confidentiality and privilege. The reason for this is that AI notetaking tools may store processed content on external servers (i.e., the cloud), and may be used to further train the AI tool and potentially become available to third parties.

In-house teams way wish to consider the following before deploying these tools across the business or within the legal function.

Privilege: What Happens When AI Is in the Room?

Even though many in-house lawyers operate within an organization context, they remain bound by the same professional duties as external counsel. That includes the duty to maintain client confidentiality and preserve privilege. Indeed, in Pritchard v. Ontario (Human Rights Commission), 2004 SCC 31, the Supreme Court of Canada reinforced that solicitor-client privilege must be “jealously guarded.”

The Law Society of Ontario’s Rules of Professional Conduct Rule 3.3-1 states: “A lawyer at all times shall hold in strict confidence all information concerning the business and affairs of a client acquired in the course of the professional relationship and shall not divulge any such information unless” certain exceptions are met. There are similar obligations imposed on lawyers across Canada by the various provincial law societies.

In the in-house setting, “client” typically refers to the organization they are representing, but may also include specific business units or executives. The duty of confidentiality and to maintain privilege extends to discussions related to legal advice or legal risk—regardless of whether litigation is anticipated.

Privilege can be waived expressly or by implication, including if shared with third parties.

If an AI notetaking tool records, stores or transmits legal advice or privileged communications—and that data is accessible to a third-party vendor or used to improve the tool’s capabilities—it could be argued that privilege has been waived. There is additional risk when the AI tool is not formally part of the company’s legal support infrastructure or when business colleagues engage these tools without understanding the implications.

Given how new some of these tools are, it is not surprising that we do not yet have specific guidance from Canada's courts on some of the key issues surrounding the waiver of privilege that may arise from the use of AI notetaking applications.

Confidentiality: Organizational Expectations

In addition to a lawyer’s obligation to maintain privilege, legal function and non-legal function employees may additionally be bound by contractual and/or common law obligations of confidentiality. Using an AI-powered notetaking tool during internal meetings (e.g., executive committee calls, incident response reviews or compliance investigations) could pose a confidentiality risk if:

• the data are transmitted or stored on servers outside the organization’s control;
• the vendor is not subject to confidentiality obligations; or
• the system uses the data to train its models or could potentially disclose the data to third parties.

Even when tools are embedded within enterprise software, the use of generative AI or transcription features should trigger a review of data handling and access protocols.

Key Risk Factors for In-House Legal Teams

In-house teams should assess AI notetaking tools through a legal risk lens, considerations include:

• Data residency and sovereignty: Where is the data stored? Are any foreign laws (e.g., U.S. CLOUD Act) applicable?
• Vendor terms of service: Do the terms permit model training, data retention or third-party sharing?
• Integration with enterprise platforms: Are tools vetted by IT and legal? Are controls configurable (e.g., disabling AI during privileged meetings)?
• Use in hybrid or external settings: How is AI handled in cross-functional meetings with consultants, auditors or regulators?

Unlike private firms, in-house lawyers must often navigate enterprise-wide adoption of technologies that are not legal-specific. Legal departments need to be proactive in defining where and how AI is appropriate, especially when meetings involve legal advice, litigation strategy or sensitive regulatory issues.

Best Practices for In-House Counsel Using AI Notetaking Tools

To manage confidentiality and privilege risk, in-house counsel considerations should include:

• Work closely with IT, procurement and privacy teams to review AI tools before adoption. Ensure the organization has a unified policy on where and how AI notetaking is used.
• Use enterprise-grade tools with clear settings to disable training and control data retention.
• Establish internal guidance on legal-use boundaries (e.g., no AI notetaking during litigation strategy sessions or compliance investigations).
• Ensure internal awareness among business stakeholders that privilege can be compromised by improper tool usage.
• Monitor developments in privacy law, AI regulation and professional conduct standards.

Takeaways: Embracing Innovation Without Losing Protection

AI notetaking tools are not inherently risky, but introduce new vectors of exposure for confidentiality and privilege, especially in complex corporate environments where legal advice is often blended with business strategy.

For in-house counsel, the goal is not to resist innovation, but to govern it responsibly. With clear policies, careful vendor evaluation and a seat at the table during tech rollouts, legal teams can ensure these tools are used safely and strategically, while protecting privilege.

If you have any questions about how your organization may use and implement AI, we invite you to contact one of the authors of this article.

Download PDF