Artificial Intelligence in the Boardroom: Governance Risks and Oversight
Artificial intelligence (AI) is beginning to make its way into boardrooms. Some companies are experimenting with software that can record meetings, generate transcripts, draft minutes, or summarize large Board packets. While these tools may seem attractive for their efficiency, they raise significant issues for corporate governance. Confidentiality, attorney–client privilege, shareholder inspection rights, SEC disclosure obligations, and fiduciary duties all come into play when Boards allow AI into their deliberations.
Confidentiality Concerns
Board meetings routinely involve the most sensitive subjects a company faces: potential acquisitions, internal investigations, forward-looking financial projections, succession planning, and executive performance. The value of these discussions lies not only in their candor but also in their confidentiality.
The use of AI-powered note-taking or transcription tools introduces new points of vulnerability. When meeting records are processed or stored outside the company’s direct control, the confidentiality of the boardroom can be compromised. For example, if an outside vendor retains a transcript of a confidential strategy session, that transcript may later be subject to subpoena, government inquiry, or shareholder demand. Even absent a legal request, any third-party storage creates exposure to cybersecurity risks.
The concern is not limited to breaches or leaks. The very existence of detailed, searchable AI-generated transcripts can make sensitive boardroom discussions more likely to be scrutinized. Plaintiffs in stockholder litigation, regulators conducting reviews, or even counterparties in negotiations may seek access to these materials to probe directors’ deliberations. What was once an informal conversation, quickly summarized in final minutes, may now be preserved in full detail and potentially discoverable.
Attorney–Client Privilege
Attorney–client privilege allows directors to receive candid legal advice during meetings, but reliance on AI tools creates new risks. If privileged exchanges are captured by an AI system and transmitted to a third-party server, courts may conclude that privilege has been waived.
AI-generated transcripts or draft minutes raise the same concern. If those records include legal advice and are processed or stored by an AI vendor, courts may treat that disclosure the same way they treat sharing privileged material with outside consultants. The result could be that sensitive legal advice, strategy discussions, or draft analyses become discoverable in litigation.
Shareholder Demands
In Delaware and other jurisdictions, stockholders are becoming more aggressive in demanding access to Board materials before filing derivative lawsuits. Courts often grant these requests when stockholders can show a credible basis to investigate possible misconduct.
This trend increases the risk that even preliminary or informal records may surface in litigation. If AI systems generate detailed transcripts or draft minutes, stockholders may argue that these documents are part of the company’s books and records and therefore subject to inspection. By creating more material and in greater detail, AI may unintentionally expand what stockholders can access, raising both the cost and the risk of litigation.
SEC Scrutiny and Disclosure Obligations
The SEC has made artificial intelligence a regulatory priority, warning companies against overstating or misrepresenting how they use it. This kind of “AI-washing” has already been called out in actions against public companies that made misleading claims about their AI capabilities.
For Boards, the risk goes beyond regulatory scrutiny. AI-generated notes or summaries may not line up with the official record. If draft minutes are mistaken for final, or if disclosure teams rely too heavily on AI outputs without proper review, important controls can break down. The result can be inaccurate filings, inconsistent communication with shareholders, and reputational or legal consequences.
The SEC has already asked companies to explain their AI use and confirm that disclosures are accurate, underscoring the need for care. The bigger point, however, is that accountability does not shift to the technology. Boards and management remain responsible for the accuracy of records and disclosures. AI may make work more efficient, but without strong oversight its use could be seen as a failure to meet the duty of care expected by both regulators and shareholders.
Fiduciary Duties and the Board Record
Directors have a duty to act with care, on an informed basis, and in good faith. When fiduciary duty claims arise, courts often look closely at the Board’s record, particularly the minutes, to decide whether directors are protected by the business judgment rule. A clear and accurate record is often the Board’s strongest defense in litigation.
AI introduces new complications to this safeguard. If AI-generated drafts fail to reflect what actually happened in the boardroom, they can be used against the company. Even small errors or omissions in AI-assisted notes can distort the record and expose directors to arguments that they acted without due care or without full deliberation.
In short, incomplete or inaccurate AI records may weaken the very protections that directors rely on in fiduciary duty litigation. What has long been the Board’s best defense, the official record, can become a source of risk unless AI outputs are carefully reviewed and verified.
Practical Guidance
Boards can capture efficiencies of AI while mitigating risks by adopting the following practices:
1. Set Clear Policies
- Define when, if at all, AI tools may be used in meetings.
- Specify that only Board-approved minutes are the authoritative record.
2. Protect Confidentiality and Privilege
- Contract with vendors for strict confidentiality, non-retention, and prohibitions on model training.
- Require agreements that preserve attorney–client privilege.
- Prohibit AI from recording privileged discussions.
- Favor company-controlled or private-cloud systems over public platforms.
3. Control and Limit AI Outputs
- Restrict AI use to draft working notes, with final minutes prepared and approved under counsel’s supervision.
- Clearly mark outputs as drafts and limit access to directors and counsel.
- Dispose of AI drafts once official minutes are approved, unless a legal hold applies.
4. Manage Litigation and Disclosure Risks
- Treat all AI-generated drafts and transcripts as potentially discoverable.
- Ensure disclosure controls account for AI outputs so filings and communications remain consistent.
- Confirm that official minutes accurately reflect reliance on expert advice and serve as the sole disclosure source.
5. Maintain Human Oversight
- Require Board secretaries and counsel to review and finalize minutes.
- Use AI only as an administrative aid, not a substitute for judgment or governance responsibility.
AI may offer Boards useful administrative support, but it also introduces risks that cut to the core of corporate governance. Confidentiality, privilege, shareholder inspection rights, disclosure accuracy, and fiduciary duties are all directly implicated. With thoughtful policies, careful vendor oversight, and strong human review, Boards can capture efficiencies while protecting the integrity of their governance. Without such measures, AI-generated records may become liabilities in shareholder litigation and regulatory inquiries.
Our Corporate and Securities Practice Group is experienced in the entire range of M&A, financings, corporate governance and related matters, and offers a sophisticated and cost-effective alternative to BigLaw rates without compromising quality. Please feel free to contact Robert L. Lawrence, rlawrence@kanekessler.com at 212-519-5103 or Jonathan A. Zalkin, jzalkin@kanekessler.com at 212-519-5195 with any questions regarding this Article or any other matter you may want to discuss.
Kane Kessler, P.C. is a full-service, general practice law firm with expertise in multiple disciplines, including Corporate, Securities and Capital Markets, Mergers and Acquisitions, Finance, Litigation, Intellectual Property, Real Estate, and Trusts and Estates.