How should directors be informed about AI use in a board evaluation?

Directors should be told, in writing and before the evaluation begins, that AI will be used, what it will analyse, what it will not decide, and how their data will be protected. Disclosure should be specific, timely and owned by the Chair, with the Company Secretary managing the process. Clear communication protects candour, which is the foundation of any credible board evaluation. The sections below set out the content, timing, ownership, consent and safeguards that make that disclosure effective.

What should directors be told about AI in an evaluation?

Directors should be told which AI tool is used, what data it processes, what role it plays in the analysis, and where human judgement takes over. Vague statements such as “we use technology to support the process” do not meet the standard that a board evaluation demands.

A complete disclosure covers five points:

  • Purpose: what the AI analyses, such as questionnaire responses, comment themes or patterns across committees.
  • Inputs: which data enters the system, including whether free-text comments and individual ratings are included.
  • Outputs: what the AI produces, for example thematic summaries, trend analysis and draft recommendations.
  • Human oversight: who reviews AI output before it reaches the board, and who owns the final conclusions.
  • Limits: what the tool does not do, including making judgements about individual director performance.

Directors respond more openly when they understand that AI supports the analysis and does not replace the evaluator’s judgement. The aim is informed participation, not technical fluency.

When should directors be informed about AI use?

Directors should be informed before the evaluation is launched, ideally when the approach is agreed and well ahead of any questionnaire being issued. Early notice allows questions, concerns and objections to surface while the process can still be adjusted.

Three moments matter most. First, include AI use in the terms of reference approved by the Chair and the board. Second, send a clear briefing note to every director when the evaluation is announced. Third, repeat the key points in the invitation to complete the questionnaire, so no director responds without knowing how the data will be handled.

Disclosure after the fact damages trust. If directors learn about AI involvement only when results arrive, they may question the validity of the findings and the integrity of the process. In a boardroom where openness and trust determine the quality of the discussion, late disclosure is a governance risk in its own right.

Who should communicate AI use to the board?

The Chair should own the message, because the evaluation is a board matter and the Chair leads it. The Company Secretary typically manages the practical communication, and the external evaluator or platform provider supplies the technical detail on request.

This division of responsibility works for a clear reason. A message from the Chair signals that AI use has board-level sponsorship and is not an administrative decision made out of sight. The Company Secretary, as custodian of governance process and records, ensures that the briefing is accurate, documented and consistent with the board’s policies on data and technology. Where the Company Secretary is closely involved in board evaluation, this role extends naturally to fielding director questions and recording any concerns raised.

The evaluator should be available to explain the methodology directly. Directors are experienced professionals and tend to ask pointed questions. Credible answers from the person accountable for the process carry more weight than a written summary.

Do directors need to consent to AI-assisted evaluations?

Formal individual consent is not always legally required, but informed agreement at board level is good governance practice. The board should approve the use of AI as part of the evaluation design, and individual directors should have a genuine opportunity to raise objections or ask for adjustments.

Legal requirements vary by jurisdiction and by the type of personal data processed. Where data protection law applies to director responses, the organisation should confirm its lawful basis with legal counsel and document it. Beyond the legal minimum, the practical test is simple: would a director feel misled if they later learned how their responses were processed? If the answer is yes, the disclosure was insufficient.

A sound approach combines three elements: board approval of the methodology, written notice to each director, and a clear route for individuals to raise concerns with the Chair or Company Secretary before they respond. If a director objects, the board can consider alternatives, such as restricting AI to aggregated data or offering a human-only review of that director’s input.

How should data privacy and confidentiality be explained?

Data privacy should be explained in plain terms: where responses are stored, who can access them, whether they are used to train AI models, how long they are retained, and how individual comments are anonymised. Directors will only speak candidly if they trust that their input is protected.

The briefing should address these questions directly:

  • Is the data processed within a defined jurisdiction, and does it leave that jurisdiction?
  • Is board data used to train or improve external AI models, or is it isolated?
  • Who sees raw responses, and who sees only aggregated results?
  • How are attributable comments handled, and can they identify an individual?
  • When is the data deleted?

Directors should also understand that confidentiality in a board evaluation is a matter of design and not assurance alone. Anonymity thresholds, role-based access and contractual restrictions on the provider should be described. The Company Secretary is well placed to verify these arrangements against the organisation’s information security and data protection policies before the briefing is issued.

How can boards address concerns about AI bias and accuracy?

Boards address bias and accuracy concerns by being candid about the limits of AI, requiring human review of every material finding, and testing outputs against the underlying evidence. AI should surface patterns that experienced evaluators then interpret, challenge and contextualise.

Three safeguards carry the most weight. First, require that every theme or recommendation can be traced to the source responses, so directors can see what drove a conclusion. Second, insist that an experienced evaluator reviews the output for context, since a pattern in the data may reflect board culture, history or a single recent event and not a systemic issue. Third, give directors the opportunity to comment on draft findings before they are finalised, which corrects errors and reinforces ownership of the result.

Directors should also hear plainly that AI can misread nuance, tone and cultural difference, particularly on multinational boards where communication styles vary. Acknowledging this does not weaken the process. It demonstrates the intellectual rigour that experienced directors expect, and it keeps the evaluation forward-looking and focused on strengthening the board’s effectiveness.

How The Board Practice helps with AI in board evaluations

The Board Practice has refined its board effectiveness methodology over more than 19 years, led by Dr. Victor Prozesky, who has been involved in board performance programmes in over 120 instances across industries and continents. From August 1, 2026, the firm’s AI-powered SaaS platform allows boards to generate or select questionnaires, complete evaluations and receive analysis with actionable recommendations, with performance tracked continuously.

Boards working with The Board Practice benefit from:

  • A process tailored in close partnership with the Chair, including how AI use is communicated to directors
  • Honest, frank feedback that identifies both strengths and development areas
  • Expert human interpretation of AI-generated analysis, grounded in experience with multicultural and multinational boards
  • Forward-looking outcomes focused on strategy, culture, dynamics and relationships, with a compliance profile strengthened as a result

To discuss how AI can be introduced into your next evaluation with transparency and director confidence, contact The Board Practice.

Related Articles