More than 88% of organizations now use AI in at least one business function, yet only 39% of Fortune 100 companies disclose any form of board-level AI oversight. That gap defines the current challenge for artificial intelligence in the boardroom: boards are expected to oversee AI strategy, risk, and compliance while also deciding how to use AI responsibly in their own governance work.
AI now has two boardroom roles. On the one hand, it serves as a governance tool to support meeting preparation, board-pack analysis, scenario modeling, and risk monitoring.
On the other hand, it’s also a subject of board oversight because management teams are deploying AI across operations, finance, customer service, product development, compliance, and decision-making.
This guide explains the practical benefits and risks of AI-assisted board governance. It also explores how AI systems can be responsibly adopted in boardroom workflows to strengthen enterprise risk management.
Key takeaways
- Boards are using AI for meeting preparation, document analysis, scenario modeling, risk monitoring, and performance benchmarking.
- Board AI oversight should cover strategy, risk, ethics, investment value, accountability, and regulatory exposure.
- A practical AI governance framework for boards should define scope, roles, risk levels, ethical principles, regulatory alignment, and review cadence.
- Boards without a formal policy can use an AI governance policy template for boards as a starting point.
- The biggest risks include weak AI literacy, biased outputs, unclear accountability, data confidentiality issues, and reduced information-sharing between management and the board.
Artificial intelligence’s dual role in the boardroom
The clearest way to understand AI board of directors oversight is to separate two ideas that are often treated as the same.
First, AI is becoming a tool that boards can use to govern more efficiently. For example, it can summarize long board packs, identify patterns across reports, compare management proposals with external benchmarks, surface relevant questions before meetings, and even support scenario planning.
Stanford researchers describe this as a shift in how boards function, process information, interact with management, and use advisors.
Second, AI is a subject that boards must oversee. That means directors need visibility into how the organization uses AI, where the highest-risk AI systems are used, who is responsible for AI-related decisions, and whether controls are sufficiently robust.
These two roles are distinct but connected. A board may use AI to improve meeting preparation while also asking whether management’s AI systems expose the company to privacy, bias, cybersecurity, regulatory, or reputational risk.
Many boards are still early in both areas. Deloitte’s 2025 boardroom research found that 66% of respondents said their boards have “limited to no knowledge or experience” with AI, while 31% said AI was not on the board agenda.
That makes AI a governance maturity issue, not only a technology issue.
How boards are using AI today
The practical use of AI in board management is more specific than broad stories about companies deploying AI across the business.
The most useful ways AI can change board processes are the ones that reduce manual preparation, improve information review, and help directors focus on higher-risk decisions.
In many cases, generative AI is used to summarize reports and draft questions, allowing directors to review large volumes of material before meetings. This is one practical example of how AI can support directors without replacing their judgment.
PwC’s 2025 Annual Corporate Directors Survey found that 35% of directors say their boards have incorporated AI and GenAI into their oversight roles. The most common uses include:
- Staying informed on trends.
- Researching peer practices.
- Evaluating performance metrics.
- Pressure-testing strategy through scenario modeling.
- Enhancing governance processes and board effectiveness.
Meeting preparation and document analysis
Board materials are getting longer, more technical, and more risk-heavy. Directors may receive hundreds of pages covering financials, legal updates, cybersecurity, regulatory developments, ESG topics, human capital, and strategic initiatives.
AI can help by:
- Summarizing board packs before meetings
- Highlighting inconsistencies across reports
- Identifying unresolved action items
- Surfacing relevant benchmarks
- Preparing suggested questions based on the agenda
- Flagging sections that require legal, financial, or risk review.
However, this does not remove the need for director review. It changes where directors spend their time. Instead of searching for issues, they can spend more time testing assumptions.
Minute-taking is another area with clear potential, although adoption remains limited. Herbert Smith Freehills Kramer reported in 2025 that 92% of GC100 poll respondents had not yet introduced AI into minute-taking processes. The same source also notes concerns about recording board discussions, data privacy, cyber risk, and the quality of AI-produced minutes.
Scenario modeling and strategic decision support
AI can support boards during high-stakes strategic discussions by testing assumptions more quickly than manual analysis. The purpose is to improve the board’s decision-making, not to turn strategic judgment over to a model.
For example, boards may use AI-assisted models to ask:
- What happens if demand falls by 15% in a key region?
- How would a regulatory change affect a proposed acquisition?
- Which customer segments are most exposed to pricing pressure?
- Where could supply-chain disruption create margin risk?
- What competitor moves could weaken the current strategy?
PwC identifies scenario modeling as one of the ways boards are already using AI and GenAI in their oversight role.
The value here is not that AI gives the “right” answer. Instead, it helps directors test more assumptions before approving strategy, capital allocation, M&A, or major transformation plans.
Risk monitoring and compliance
AI can also help boards monitor fast-moving risk areas. This includes regulatory change, litigation trends, transaction patterns, internal control exceptions, cybersecurity signals, and compliance obligations.
For example, an audit or risk committee may leverage AI-assisted reporting to identify:
- Unusual transaction behavior
- Emerging regulatory requirements
- Gaps between policy and actual practice
- Vendor-related risk concentration
- Unresolved compliance issues across business units.
This is where AI tools for corporate governance can support board effectiveness analysis when they are connected to reliable data, clear escalation rules, and human review. Stanford’s 2025 boardroom paper also notes that AI can support governance functions, including strategy, audit, legal monitoring, human capital oversight, compensation analysis, and board evaluations.
Still, the board should evaluate whether AI improves governance quality or simply adds another layer of complexity.
Reducing information asymmetry between boards and management
AI may also change one of the oldest tensions in corporate governance: boards depend heavily on the information management chooses to prepare.
Stanford researchers argue that AI can give directors more access to deeper information beyond management-prepared board materials. It may also prompt directors with questions related to the agenda, suggest analyses that support decision-making, and make boards more proactive.
It may also change the role of board advisors, who can help directors interpret AI-assisted analysis rather than simply deliver static reports.
However, this comes with a new risk. A 2025 ECGI working paper by Daniel Ferreira and Jin Li argues that AI can also serve as a private advisor to CEOs, reducing management’s incentive to share information with directors. The authors conclude that board adoption of AI may reduce some negative effects of CEO AI use, but cannot fully restore efficient board monitoring.
For boards, AI should improve information flow without creating a new layer of hidden management analysis.
The board’s AI oversight responsibilities
Boards have always overseen strategy, risk, leadership, controls, and long-term value. AI does not create a separate universe of director duties. It adds a new layer to existing fiduciary responsibilities.
The urgency is clear. McKinsey reports that only 39% of Fortune 100 companies disclosed any form of board oversight of AI as of 2024, even as enterprise adoption has accelerated.
Strong board AI oversight usually includes five responsibilities:
- Keep AI on the board agenda regularly. A single annual AI presentation is usually not enough for companies where AI affects operations, customer experience, compliance, or competitive position. Deloitte’s 2025 survey found that only 17% of respondents said AI is addressed at every board meeting, while 19% said it comes up only once per year.
- Build AI literacy across the board. Directors need enough knowledge to understand where AI affects strategy, risk, compliance, and stakeholders. This does not require coding knowledge, but it does require confidence in asking management the right questions.
- Clarify who owns AI oversight. The board should decide whether AI oversight sits with the full board, the audit committee, the risk committee, the technology committee, or a dedicated AI group. Deloitte notes that boards may support AI governance through a dedicated subcommittee or by expanding an existing committee’s mandate.
- Connect AI investment to measurable business value. Directors should ask how management measures AI value. Cost reduction, revenue growth, productivity, risk reduction, compliance improvement, and customer outcomes may all matter, but the board needs clear metrics.
- Ensure responsible use in board processes and risk controls. AI risk management for boards should include bias testing, privacy controls, cybersecurity safeguards, vendor due diligence, human review, auditability, and escalation procedures. Directors should also ask how AI is being used across internal operations, especially in finance, HR, compliance, customer service, and risk management.
For a broader context on board duties, see our guide on the importance of corporate governance
Building a board AI governance framework
A practical AI governance framework gives directors a structured way to oversee AI use without turning every board meeting into a technical review.
The goal is to define how AI decisions are made, who is accountable, which systems require board visibility, and how risks are reviewed over time. For many organizations, this starts with a formal AI governance framework for boards.
At a minimum, the framework should cover six areas.
1. Scope
Define which systems are covered, including internally developed models, third-party AI tools, generative AI applications, automated decision systems, AI-enabled analytics platforms, and tools used in board administration.
The scope should also clarify whether the policy applies only to enterprise AI or also to generative AI in the boardroom, such as summarization, agenda support, board-pack review, and draft minutes.
2. Roles and accountability
The framework should explain who owns AI decisions at the management level and how the board receives reporting. Typical accountability questions include:
- Who approves high-risk AI use cases?
- Which executive owns AI risk?
- Which committee receives AI updates?
- When does an AI issue escalate to the full board?
- Who validates vendor controls?
- Who confirms that AI outputs are reviewed by humans?
That is why board member responsibilities on AI should be made practical and easy to understand.
3. Risk classification
Not every AI use case carries the same risk. A tool used to summarize public market research is different from a model used in credit decisions, hiring, pricing, healthcare recommendations, insurance underwriting, or legal analysis.
A board-level framework should classify AI systems by risk level. Higher-risk systems should receive stronger documentation, testing, monitoring, and escalation.
4. Ethical oversight
Responsible AI governance should include principles for fairness, transparency, accountability, security, privacy, explainability, and human oversight.
Therefore, a board should ask management to explain how they test bias, monitor model performance, correct errors, and allow affected stakeholders to challenge or escalate AI-supported decisions.
5. Regulatory alignment
The regulatory environment is changing quickly. The EU AI Act entered into force on August 1, 2024, with phased application dates. AI literacy obligations began applying from February 2, 2025; governance rules and general-purpose AI obligations became applicable on August 2, 2025, and certain high-risk AI rules have extended timelines.
Even companies outside the EU may feel the effect if they operate in European markets, serve EU customers, or use vendors subject to the Act. Boards should also consider sector-specific rules, privacy laws, securities disclosures, employment laws, consumer protection rules, and industry guidance.
6. Review cadence
Boards should review the framework at least annually. For organizations with high AI exposure, a quarterly committee-level review may be more appropriate.
For a broader view of how boards oversee technology, data, and digital risk, see our guide on digital governance
Challenges and associated risks of AI in boardroom governance
The risks of AI in boardroom governance affect director preparedness, data security, fiduciary oversight, regulatory exposure, and the quality of board-management dialogue.
AI literacy gap on the board
Boards cannot oversee what they do not understand.
This gap can show up in several ways:
- Directors accept management’s AI claims too easily
- Committees cannot distinguish strategic AI from isolated automation
- Risks are discussed too late
- AI investments are approved without clear metrics
- Vendors are assessed mainly on functionality, not governance controls.
Deloitte recommends building board AI literacy through subject-matter experts, independent learning, and direct engagement with management AI leaders.
Bias and fairness in AI decision support
Bias has become a board-level governance concern.
AI systems can reflect biased training data, weak assumptions, incomplete inputs, or flawed proxy variables. If management uses these outputs in pricing, hiring, lending, customer segmentation, insurance, healthcare, or compliance decisions, biased recommendations may create legal risks and reputational harm.
Boards should ask:
- How does management test AI outputs for bias?
- Which stakeholder groups could be affected?
- Who reviews high-impact recommendations before action is taken?
- How are errors documented and corrected?
- Are external vendors required to disclose testing methods?
Human review matters most when AI influences decisions that affect people, access, opportunity, pricing, or rights.
If your board is reviewing AI through a wider risk and stakeholder lens, see our guide on ESG and board of directors
Transparency, accountability, and the black-box problem
Many AI systems are difficult to explain in plain language. That creates a real challenge in a governance setting: boards may be asked to oversee decisions whose logic is unclear.
The EU AI Act includes transparency obligations for certain AI systems and GPAI models, while high-risk systems require governance controls, including documentation, human oversight, and monitoring.
For boards, explainability does not mean every director must understand model architecture. The real test is whether current directors are equipped to ask management for clear, decision-ready explanations.
At a minimum, management should be able to explain:
- What the AI system does
- What data it uses
- Where it is used
- What decisions it influences
- What its known limitations are
- Who can override or challenge the output
- How performance is monitored over time.
These points should help directors raise important questions before AI-supported recommendations influence strategy, reporting, compliance, or stakeholder outcomes.
A board should be uncomfortable with any high-risk AI system that management cannot explain at the level of oversight required.
Data confidentiality and the boardroom privacy risk
Board materials often contain highly sensitive information: M&A plans, executive compensation, litigation updates, cyber incidents, financial forecasts, regulatory investigations, restructuring plans, and succession discussions.
That makes consumer AI tools risky for board work. If directors or governance teams paste confidential board materials into unapproved AI tools, they may expose privileged, personal, or commercially sensitive information.
Herbert Smith Freehills Kramer notes that AI use in board-related tasks raises issues around board discussion quality, cyber risk, voice recordings, and data privacy.
Boards need clear rules on:
- Which AI tools are approved
- Whether board materials can be uploaded
- Whether meeting recordings are permitted
- How prompts and outputs are stored
- Whether AI-generated summaries become board records
- Who reviews AI-produced minutes before approval.
A thoughtful approach allows directors to use AI-generated insights where they add value, while keeping confidential board materials within approved systems, clear policies, and secure governance workflows.
The agency problem: AI reducing board independence
AI can reduce information asymmetry between boards and management. It can also deepen it.
The Ferreira and Li working paper explains that AI may act as a private advisor to CEOs, reducing their incentive to seek board advice or share information. In that scenario, management becomes more self-sufficient, while directors receive less visibility into how strategic decisions were made.
This creates a new governance question: is AI improving board oversight, or is it helping management make more decisions before the board can challenge the assumptions?
Boards can reduce this risk by requiring management to disclose:
- Where AI supported major recommendations
- Which assumptions were used
- What alternatives were rejected
- Where human judgment changed the AI output
- Which risks remain unresolved.
AI should strengthen the board-management relationship by improving the quality of discussion. It should not become a private management layer that directors cannot see.
Building AI literacy on the board
AI education for board members should not be overly technical. Directors need sufficient AI governance fluency to understand how AI impacts strategy, risk, oversight, and accountability.
Three approaches work especially well.
- Bring in subject matter experts for board education sessions. These can include internal technology leaders, external AI governance advisors, legal counsel, cybersecurity experts, and sector specialists. External board advisors contribute the most value when they translate AI concepts into practical questions directors can use in strategy, risk, and compliance discussions.
- Review board composition. Deloitte’s 2025 survey found that 40% of respondents are rethinking board composition because of AI. This does not mean every board needs a full-time AI scientist, but nomination committees should consider whether the board has enough digital, data, cybersecurity, risk, and technology oversight experience.
- Define the committee structure. Some organizations may create an AI-focused board subcommittee or designate a mandate for a committee. Others may expand the mandate of the audit, risk, technology, or governance committee. The right answer depends on the company’s size, industry, risk profile, and AI maturity.
To assess whether your board has the right skills for AI oversight, see our guide on board evaluations and self-assessments
Questions AI-literate directors should ask
Strong AI board literacy shows up in better questions:
- What AI systems are already in use?
- Which use cases affect customers, employees, investors, or regulators?
- What controls exist for high-risk systems?
- How does management measure value?
- What happens when AI outputs are wrong?
- Who is accountable for model performance and human review?
That level of literacy is enough to support disciplined board oversight of AI strategy.
How Ideals Board supports AI-ready governance
A board navigating AI needs secure information flows, clear agendas, controlled access to sensitive materials, and reliable records. A board portal with AI can support AI-enhanced board collaboration while keeping sensitive governance materials in a controlled environment.
Ideals Board supports AI-ready governance by helping boards organize the documents, meetings, and workflows around AI oversight.
Here’s how a board portal improves efficiency and saves time:
- Secure document repository. Store AI policies, board materials, committee reports, and risk updates in one controlled environment.
- Granular access permissions. Control who can view, download, or manage sensitive AI-related materials.
- Meeting management tools. Schedule AI education sessions, committee reviews, and recurring oversight discussions.
- Agenda creation and board-pack organization. Structure board agendas around AI strategy, risk, compliance, and policy updates.
- Minutes and decision records. Document AI-related discussions, approvals, action items, and follow-ups.
- Task tracking and follow-up management. Assign next steps after AI oversight discussions and track progress between meetings.
- AI-enabled board portal capabilities. Support secure, controlled board collaboration around AI-related materials and workflows.
Ideals Board also offers an AI board portal page with more details on AI-enabled board management capabilities.
