Over the past two years, the launch of generative artificial intelligence (generative AI) has stood out as a potentially transformative tool for individuals and organizations. As generative AI becomes more common in organizations, much has been written about the role the board of directors should play to enable its capabilities while overseeing its responsible use. But what about the board’s own use of generative AI? What role can generative AI play in advancing the work of directors and management to drive effective functioning and stronger governance practices?

Just as boards need to be knowledgeable enough to govern organizations that use AI, they also need to learn to use these tools themselves. It is healthy for boards and the people who work with them to thoughtfully consider how, when, and why to use AI to enable the board’s own work – and the questions its use raises.

At Watson, we are exploring and envisioning how boards of directors, senior management, and governance professionals will harness generative AI to make their collective work stronger. In a world of intersecting risks and new opportunities for organizations, is generative AI a powerful tool to elevate effective governance, or does it raise more concerns and risks than benefits?

To understand the implications of generative AI on boards and management teams, let’s consider a scenario:

Fast-Forward to 2028: Generative AI in Action for a Committee Meeting

Imagine it’s 2028. Generative AI has become a seamless part of how businesses – and boards – do their work.

In preparation for an upcoming board of directors meeting, a CHRO is preparing their general report on HR metrics and matters for the Human Resources Committee (HRC). Instead of starting from scratch, they use AI to streamline the process:

• Summarize and Ideate: The CHRO asks their generative AI tool for a summary of the minutes and action items raised by the HRC at the last meeting, and suggestions for addressing each action item to ensure they have been adequately responded to.
• Drafting Support: Using tailored prompts with key topics in mind, they ask the generative AI-supported features of their enterprise systems to produce a rough draft of their report, complete with charts and figures automatically populated from internal data that the tools have been granted access to.
• Instant Insights: Not wanting to miss key updates, they use their in-house licensed AI assistant to scan through their emails and chats from the last quarter to surface key narrative points and updates to integrate into the report.
• Edit and Refine: After stepping back and reviewing the report with a human eye to check that the report is coherent and has the audience in mind, they ask the generative AI assistant to suggest edits for clarity and readability before generating a briefing note that summarize the most important points and questions.

Ready to go? They load it into the board portal.

A few days later, an HRC member accesses the board portal. They review the report and note their questions and concerns. Not wanting to miss anything, they ask the built-in generative AI assistant some questions:

• Understanding Trends – How are key KPIs trending over the last 12 months? What questions can we ask to drill into unexpected changes?
• Variance Analysis – Do the HR budget items align with the approved budget? Is there coherence and alignment from Committee to Committee?
• Industry Insights – Have best practices been adequately considered in the HR strategy and roadmap? What would best in class look like in this sector?
• Risk Analysis – Are there unseen risks in the proposed new HR system upgrade? What are the most common pitfalls of HR system transformation? Do we consider this in our risk register?
• Diligence Check – What stands out as the most important issues and questions? What questions should we be asking and might be missing?

Expanding on the Generative AI Applications for Boards of Directors

This scenario above only scratches the surface; there are so many potential applications, many of which we can reasonably expect to come to life over the next several years.

Aside from meeting preparation, imagine the role generative AI can play in:

• Effective meetings: Supporting the Chair to manage the agenda and time, soliciting input from others, summarizing discussions, and providing alternative or competing views for consideration.
• Board composition planning: Identifying emerging areas of expertise for the board to consider and conducting initial research and outreach to potentially suitable candidates.
• Director and chair feedback: Analyzing meeting transcripts, director peer feedback surveys, and psychometric assessments to provide insights into director effectiveness and areas for improvement.
• Scenario planning and analysis: Using generative AI tools to provide and evaluate potential risks and strategic decisions for the board to consider.
• Benchmarking: Comparison to peer groups in key organizational performance indicators or executive compensation.

The potential is huge – and it is also timely. As boards navigate an increasingly complex and demanding word, they will need tools and capabilities to ensure they are engaged and focused on the right issues at the right time.

Navigating Generative AI-Related Uncertainty

While we see the potential, there are many risks and questions to consider, including trust in the underlying technologies, access to accurate and integrated data across the organization, protection of personal and confidential information, ethical implications, misinformation, the environmental impact of such tools, and more.

Balancing both the opportunities and potential risks and implications, here are some of the questions we have been asking ourselves:

Benefits vs. Limitations

• Given that generative AI is trained on specific sources of data, will using it tend to limit ideas or rely heavily on past practice, failing to consider the novel or differentiated approaches required to outperform others?
• Generative AI is fallible, especially when dealing with emerging or complex issues. Will the technologies advance to a point where generative AI can be an effective tool in tackling these situations, or will using it limit our accurate understanding of them? Will alternative approaches such as Retrieval Augmented Generation or Small Language Models address this concern?

The Trust Dilemma

• Directors have a duty of care. In exercising that duty, there is a principle of reasonable reliance around what information they trust and use to make their decisions. What does reasonable reliance look like when information is sourced or reviewed with the help of generative AI? Beyond information, what about advice or recommendations that are sourced from generative AI? What legal or ethical risks arise in these situations?
• When directors receive information or advice, they consciously or unconsciously consider the source. Whether it’s an executive, advisor, a trusted news source, a systems-generated report, or a research paper, directors attach credibility to what they are receiving based on where it comes from. How can they do so with generative AI-sourced insight if the underlying sources are unclear?

Diversity and Bias – How Does Generative AI Help or Hurt Us?

• Researchers have demonstrated that decisions improve when decision-making groups are diverse. Part of this is attributable to having people who think differently and bring different perspectives or bases of experience, which can help avoid group think and limit biases. Does generative AI increase the diversity of thought at a board table, or is it trained on a narrower pattern of information that may hinder future progress?
• When might generative AI help us interrogate our assumptions and biases? When might it introduce biases? When might it simply mirror our existing biases?
• Does any of the above change based on what the topic or type of decision is? How can directors envision which topics are appropriate to involve generative AI, and the degree to which it should be considered in making a decision?

Future-Ready Directors

• How do directors develop the critical thinking and competencies to navigate generative AI and use it wisely in the board context? How does it change the role and competencies of a director?
• If generative AI proves to be a useful timesaver and efficiently support analysis and decision making, how will directors and management elevate their roles to continue having positive impact at the boardroom table?

As we improve our understanding of generative AI, gain more collective experience using it, and watch for potential legal and business tests of its use, these are the issues boards will need to tackle. We will experience the benefits and risks associated with generative AI firsthand and we’ll refine our views, as will our regulators and service providers. Boards who embrace generative AI thoughtfully and judiciously, even in small, experimental ways, will position themselves to lead their organizations with positive impact. This could include assessing everyone’s understanding and comfort level with generative AI, hands-on training to inform initial experimentation with approved tools and information, and discussing potential uses within the organization or to support the board’s work at meetings.

By strengthening understanding of generative AI and its potential benefits and pitfalls, directors will sharpen their oversight of how their respective organizations use this technology and concurrently shape our collective understanding of its role in the boardroom. We don’t presume to have the answers, but these are the questions we’re working on, and we invite you to join the conversation.