Leading the Charge: Considerations for an AI Governance Framework

Artificial intelligence has been used for decades but has grown in popularity since OpenAI launched ChatGPT in late 2022. Since then, investors and the public have become enamored with AI tools, pushing corporate issuers to follow, adopt and disclose their use of these tools.

On one hand, AI provides numerous opportunities for organizations to drive innovation and increase productivity. On the other hand, AI carries significant risks that organizations must manage.

Boards of directors and management teams, for example, are expected to adapt to changing expectations and the evolving technological landscape to stay ahead in an increasingly competitive environment. There are many reasons why it is important to have processes in place to appropriately oversee the development, use and management of AI systems and technologies. Investors’ views, ethical concerns and legal/regulatory developments are a few considerations that drive AI governance adoption.

Boards are increasingly using and educating themselves on AI. However, since AI is a relatively new and rapidly evolving board consideration, the degree to which organizations have integrated and disclosed the board’s oversight varies.

Generally, few public companies disclose details of their board’s oversight of AI, but those that do, approach the oversight of AI in a variety of ways. Some public companies note that the full board is responsible for AI oversight. Others delegate this responsibility to a specific committee, such as the audit or risk committee, or across multiple committees. We anticipate that AI-related disclosure will continue to increase during this proxy season.

As part of its 2025 proxy voting guidelines, Glass Lewis introduced guidelines regarding the board’s oversight of AI. Glass Lewis will not typically make voting recommendations against the election of directors for the organization’s AI related disclosure. In cases where the organization’s failure to oversee AI results in material harm to shareholders, Glass Lewis may consider recommending a vote against the directors responsible for this oversight. Institutional Shareholder Services has not yet adopted a comparable voting recommendation on the board’s oversight of AI.

Some TSX-listed companies have already disclosed various practices they implemented regarding their use of AI. For example, some have adopted principles for the ethical use of AI, specific policies governing AI usage within the organization, and AI governance frameworks.

When considering the adoption of an AI governance framework, organizations should ensure that it promotes transparency and knowledge throughout the organization, while limiting risks that may ensue from AI use. At a high level, an AI governance framework should cover the following:

• Responsibilities – The framework should identify who (the full board, committee, director and which member of management) is responsible for AI oversight at the organization, and what the responsibility entails. If it is a joint responsibility, the organization should clearly identify which responsibilities each group is responsible for. Reporting mechanisms should also be set out in the framework.
• Awareness – One aspect of the oversight role is to understand how AI is used and how it may be used at the organization.
• Policies and Processes – Given the risks in using AI, the board and management should work closely to put policies and processes in place that promote transparency, ethical use, accountability and oversight of AI at the organization.
• Communication and Education – AI policies must be communicated throughout the organization. Employees should be educated on appropriate use of AI, the limitations associated with its use and the expectations of the organization’s decision-makers regarding AI use and disclosure.
• Engagement – Investors care about the processes of the organization and, so, the board and management should communicate their strategies to investors and other stakeholders. The organization will also benefit from receiving input from investors on their expectations or even their use of the AI processes (if some systems are investor facing). Additionally, engagement sessions with other stakeholders such as employees may also shed light on internal processes not working efficiently.
• Continued Review and Reflection – The board and management should stay abreast of new developments in the area. Board education sessions, for example, are an appropriate way to stay up to date on new developments.

April 1, 2025