Orchestrating Oversight
If only one committee owns artificial intelligence, then nobody owns it. Effective oversight distributes the work deliberately and names someone to hold it together.
In Brief
The Problem. Boards facing artificial intelligence (AI) ask "Who owns this?" and often hand it to one committee. AI runs through clinical care, finance, compliance, workforce, and reputation, so a single owner creates a single point of failure and lets the rest of the board look away.
Why It Persists. Most boards discuss AI but have not assigned it. Few have written AI oversight into committee charters, and the hardest questions, such as approving a clinical AI tool, cross committee lines that no charter covers.
What Boards Should Do. Map each AI oversight duty to the committee with the most relevant expertise, write primary and secondary responsibility into the charters, define how committees work together at the intersections, and make the chair or lead independent director responsible for orchestration.
When AI first appears on a board agenda, the first question is usually "Who owns this?" The instinct to assign this complex technology to a single committee, such as Audit or a newly formed Science and Technology Committee, is understandable. The approach is nonetheless flawed. AI is not a single issue to be isolated. It is a horizontal capability that runs through every part of the enterprise, from financial risk and talent strategy to clinical safety and ethical reputation. Effective governance therefore requires a deliberate, orchestrated distribution of responsibilities across the existing committee structure.
What Boards Are Doing Now
Most boards have started the conversation but not the structure. The National Association of Corporate Directors (NACD) surveyed public company boards in 2025. Sixty-two percent had set aside full-board agenda time for AI. Only 27 percent had written AI oversight into committee charters, and six percent had formed a dedicated technology committee (National Association of Corporate Directors, 2025). The largest companies are moving faster. In 2025, about 40 percent of Fortune 100 companies disclosed that at least one board committee oversees AI, up from 11 percent a year earlier. The audit committee is the usual choice. Companies that assign AI to a technology committee or to a nominating and governance committee describe the duty in more detail and are more likely to write it into the charter (Niemann, 2025).
Many boards have not yet begun. In the PwC 2026 survey of about 600 public company directors, 60 percent said AI is not yet part of board oversight at all (Gaskin, 2026). Discussion without assignment leaves each committee to assume that another is watching.
Why One Committee Is Not Enough
The temptation to assign a single owner overlooks the many sides of AI. A one-committee solution creates a single point of failure and risks marginalizing the issue. It also sends the wrong signal to the rest of the board: that they need not develop their own literacy on the subject.
The law points the same way. Under In re Caremark International Inc. Derivative Litigation, 698 A.2d 959 (Del. Ch. 1996), every director shares the duty to make a good-faith effort to ensure that a reporting system exists. In Marchand v. Barnhill, 212 A.3d 805 (Del. 2019), the Delaware Supreme Court applied that duty with particular force to "mission critical" risks (Zauderer et al., 2023). AI used in diagnosis, claims, or trial evidence meets that description.
Committee structure itself can become the evidence. In the Teligent case, decided on September 2, 2025, the Court of Chancery let oversight claims proceed against a drug maker's directors. The complaint alleged that no board committee was responsible for compliance with Food and Drug Administration requirements for a prolonged period (Bonadurer, 2026). A gap in the charters is not an administrative detail. It can be the first fact a plaintiff cites.
Assign Primary Duties by Expertise
A sounder approach maps each AI governance duty to the committee with the most relevant expertise, and then writes the assignment into the charter. The Audit and Risk committees are the natural stewards of AI risk management and investment oversight, because of their expertise in financial controls and systemic threats. The Science and Technology Committee is best placed to examine whether algorithms work as claimed. The Governance and Compensation committees should take primary responsibility for the board's own AI education and for the talent strategy the organization needs to compete.
Health systems and life sciences companies use different committee structures, so the same duties land in different places. The table shows a typical assignment.
| AI Oversight Duty | Life Sciences Company | Health System |
|---|---|---|
| AI risk in the risk register, internal controls, and assurance over the AI inventory | Audit, or Audit and Risk | Audit and Compliance |
| Investment case and measures of return for material deployments | Audit and Risk, or Finance where one exists | Finance |
| Technical validation and performance monitoring after deployment | Science and Technology | Quality and Patient Safety |
| Ethical scrutiny, including bias and patient impact | Science and Technology, with an ethics committee where one exists | Quality and Patient Safety, with the ethics committee |
| Board AI literacy, director recruitment, and the allocation of AI duties | Nominating and Governance | Governance |
| AI talent strategy and executive incentives | Compensation | Compensation |
The principle is the same in both. The charter, not habit, says who looks at what.
The Work Happens at the Intersections
Distributing primary responsibility is the first step. The most complex challenges, such as the ethical application of clinical AI, do not fit neatly into one committee's charter. The oversight of a new diagnostic tool, for instance, requires three perspectives. It needs technical validation from the Science and Technology Committee (or Quality and Patient Safety in a health system), ethical scrutiny from an ethics committee, and the risk perspective of the Audit Committee. Without defined protocols for collaboration, important issues fall into the gaps between committee meetings.
External standards show how wide the intersection is. In September 2025, the Joint Commission and the Coalition for Health AI issued guidance on the responsible use of AI in healthcare. Its seven elements run from policies and governance structures and data security to quality monitoring, risk and bias assessment, and training, and it calls for multidisciplinary oversight and regular updates to senior leadership and the governing board (Polsinelli, 2025). Those seven elements span at least four committees. The Joint Commission's voluntary Responsible Use of AI in Healthcare certification, launched on June 1, 2026, asks hospitals to show governance, safeguards, monitoring, and staff education (Twenter, 2026). Preparing for it requires exactly this kind of coordination.
Three practices close the gap. First, each charter names primary and secondary responsibility for AI, so overlap is intentional rather than accidental. Second, committee chairs exchange a short written AI report each quarter, so each committee sees what the others found. Third, the board holds a joint session when a deployment touches several committees. A clinical AI tool that also changes billing is the common example.
When a Dedicated Committee Makes Sense
This reality has led some organizations to consider an AI and digital governance committee. A dedicated committee can serve as a central hub, provide specialized expertise, and coordinate the work of the other committees. It offers deep focus, but it also presents a risk: the other committees may abdicate their responsibilities and treat AI as someone else's problem.
The decision to create a new committee is a strategic one that depends on the organization's context, maturity, and risk profile. For a company where AI is central to the business model, such as a developer of AI-enabled devices or a drug developer whose discovery platform is built on AI, a dedicated committee may be essential. For most health systems, clear charters and a coordinating mechanism are enough. Either way, a dedicated committee should coordinate the duties of the others, not absorb them.
The Strongest Objection: Shared Ownership Is No Ownership
The strongest objection to a distributed model is the one this article makes against a single committee, turned around. If six committees each own part of AI, each can point to the others when something fails. Accountability that is spread thin can disappear. A single committee, on this view, at least gives the board one place to look.
The objection is right about unmanaged distribution and wrong about the remedy. Concentration does not remove the problem; it moves it to the boundary between the AI committee and everyone else. The answer is orchestration with names attached. Each charter states primary and secondary responsibility. One executive is accountable to the board for the AI inventory and reports in a common format to every committee that uses it. One director, the chair or the lead independent director, answers for the whole. With those three elements in place, distributed ownership is easier to audit than concentrated ownership, because each committee's minutes show what it reviewed.
The Chair Holds It Together
Whether a board establishes a new committee or relies on its existing structure, the mandate is the same: orchestration. The board chair and the lead independent director must ensure that the committees operate as an integrated system rather than as independent silos. Orchestration also requires confirming that the minutes show each committee did its part.
The goal is a federated governance model, in which distributed ownership is unified by a shared inventory, a common reporting format, and a named executive accountable to the board. The federated model turns the committee structure from a source of confusion into a source of clarity, so that no part of AI is left without oversight.
I have found that the annual charter review is the simplest place to start. It is already on the calendar, it requires a vote, and it leaves a record.
For the Board
- Which committee charter names AI oversight today, and which committees assume another is responsible?
- Do our charters state primary and secondary responsibility for each AI duty, including ethical review?
- Who is the executive accountable to the board for the AI inventory, and how often does that person report?
- When a deployment touches several committees, what triggers a joint review?
- Do committee chairs see one another's AI findings before the full board meets?
- If we pursue the Joint Commission certification, which committee owns the decision and the evidence?
Find the question sets for each committee seat in the Avancer AI Governance Navigator.
References
Bonadurer, J. (2026, February). Delaware court denies dismissal of Caremark oversight claims in Teligent case, highlighting board and officer compliance obligations. Jenner and Block. https://www.jenner.com/en/news-insights/publications/delaware-court-denies-dismissal-of-caremark-oversight-claims-in-teligent-case-highlighting-board-and-officer-compliance-obligations
Gaskin, J. L. (2026, September 23). Boards say they need more from management on AI. Corporate Compliance Insights. https://www.corporatecomplianceinsights.com/boards-say-they-need-more-from-management-ai/
National Association of Corporate Directors. (2025). AI oversight activities conducted by the board [Data table]. 2025 Public Company Board Practices and Oversight Survey. Retrieved October 21, 2026, from https://www.nacdonline.org/all-governance/governance-resources/governance-surveys/surveys-benchmarking/bpo-tables/ai-oversight-activities-conducted-by-the-board
Niemann, P. (2025, October 28). Cyber and AI oversight disclosures: What companies shared in 2025. Harvard Law School Forum on Corporate Governance. https://corpgov.law.harvard.edu/2025/10/28/cyber-and-ai-oversight-disclosures-what-companies-shared-in-2025
Polsinelli. (2025, October 29). Joint Commission and Coalition for Health AI release first-of-its-kind guidance on responsible AI use in healthcare. https://www.polsinelli.com/publications/tjc-chai-release-guidance-responsible-ai-use-healthcare
Twenter, P. (2026, June 1). Joint Commission launches certification for responsible healthcare AI use. Becker's Hospital Review. https://www.beckershospitalreview.com/quality/joint-commission-launches-certification-for-responsible-healthcare-ai-use/
Zauderer, F. M., Monteleone, J. P., & Fenichel, A. H. (2023, August). Guest post: Board oversight duties, recent adventures in the Delaware Chancery. The D&O Diary. https://www.dandodiary.com/2023/08/articles/director-and-officer-liability/guest-post-board-oversight-duties-recent-adventures-in-the-delaware-chancery/