Boards can govern the ethical use of AI in human resources effectively by insisting on three things at once: formally adopted ethical principles, a comprehensive written policy, and a cross-functional oversight committee with real decision rights. Guidance for boards and risk leaders from Diligent, together with a published corporate AI governance policy from HR Acuity, converges on this structure. The board's job is not to audit algorithms itself, but to verify that management has built the machinery — principles, policy, committee, monitoring and reporting — that makes responsible AI adoption in recruitment and performance assessment verifiable rather than assumed.
Pillar 1: Establishing Ethical Principles for HR AI
Every governance structure rests on a small set of ethical principles that the organization commits to in writing. Two in particular carry the weight in HR contexts, where decisions affect people's livelihoods directly: fairness and transparency. A third, accountability, binds the first two to named owners.
Fairness in this context means AI systems are designed and used to produce equitable, unbiased outcomes, actively avoiding discrimination in HR processes. HR Acuity's published policy describes its approach as "designed to minimize the risk of discriminatory outcomes against any group or individual," achieved through what it calls thoughtful implementation, thorough audits and comprehensive adjustments to its AI features and related practices. The critical qualifier is that fairness is not a one-time design property: it requires continuous monitoring and mitigation of potential biases in both data and algorithms.
Transparency means the operation and decision-making processes of AI systems should be understandable and explainable to relevant stakeholders — especially in HR applications, where candidates and employees have a legitimate interest in knowing whether and how a machine influenced a hiring or performance decision. HR Acuity's policy commits the company to sharing information about its AI usage with end-users and ensuring that end-users are not misled into believing AI-generated outputs are human-generated. The degree of transparency may vary with system complexity and proprietary considerations, but the rationale behind decisions should remain clear.
Accountability closes the loop: clear lines of responsibility must exist for AI system development, deployment and outcomes within the HR function. In HR Acuity's framing, accountability extends to ensuring adherence to the ethical pillars and monitoring AI usage for compliance. For a board, the practical test is simple — if a biased screening outcome surfaces next quarter, can management name the person or body responsible for catching it?
Pillar 3: Implementing Effective Oversight Mechanisms
Principles without oversight decay into marketing language. Diligent's guidance for boards, risk and audit leaders, authored by Kezia Farnham, identifies the components that turn principles into a functioning policy: a defined scope covering which AI systems the policy applies to (in-house, procured or embedded; models, automated decisions and analytics); governance principles including fairness and bias mitigation, transparency and explainability, accountability and oversight, privacy and data protection, and security and risk management; explicit alignment to compliance and legal frameworks such as the EU AI Act, the NIST AI Risk Management Framework and the OECD AI Principles; defined roles and responsibilities for who owns governance, model standards, legal and risk review, and user reporting; and a set cadence for AI risk assessments and audits, including when third-party audits apply to high-risk systems.
Those policy components matter to boards because HR AI spans the full employee lifecycle — recruitment screening, performance assessment, workforce analytics — and each use case carries different risk. A policy that covers only in-house models while ignoring AI embedded in a vendor's applicant tracking system leaves the highest-exposure gap unaddressed.
Oversight then has to be operational. Diligent recommends moving the AI inventory, risk register, control library and board reporting off spreadsheets and onto a single platform with approval workflows and risk heatmaps, warning that document-based governance cannot keep pace with AI adoption. The final step is continuous: monitor, audit and improve. HR Acuity's policy assigns its governing body the duties of monitoring usage, conducting audits and identifying anomalies, with the body meeting periodically to review and revise policies, audit processes and ensure adherence to principles — and holding authority over disciplinary actions for non-adherence.
Board-level reporting is the mechanism that connects all of this to the boardroom. Diligent's guidance treats regular board reporting as part of operationalizing governance, and reporting should cover adherence to the ethical pillars, usage audits and the identification of anomalies. Without a recurring report on AI risks and compliance, a board is governing on assertion rather than evidence.
Establishing and Empowering the AI Ethics and Compliance Committee
The operational heart of the framework is a dedicated AI ethics and compliance committee. Diligent is explicit about composition: cross-functional by design, drawing on risk, compliance, legal, IT, security, HR and at least one business representative. The breadth matters in HR contexts specifically — legal and compliance expertise is needed for employment-law exposure, HR for the talent processes themselves, and IT and security for the systems and data involved.
Composition without authority is the most common failure mode. Diligent flags the pitfall directly: a committee with no decision rights. Its remedy is to give the committee authority over new-use-case review and training. In HR terms, that means no new AI-assisted recruitment or performance tool enters the organization without the committee's review, and the committee controls the training that prepares employees to use AI responsibly.
HR Acuity's own governing body illustrates the mandate in practice. Its responsibilities include establishing overarching accountability within the AI strategy, ensuring strict adherence to the AI Ethics Pillars, monitoring usage, conducting audits and identifying anomalies, and ensuring AI technologies integrate with existing infrastructure to enable operational efficiencies. The body meets periodically to review and revise policies, audit processes and adherence to principles, and it carries responsibility for disciplinary actions when the policy is breached. That last element — consequence authority — is what distinguishes a governing body from an advisory group.
AI Governance Framework for Board Oversight in HR
The matrix below synthesizes the expert guidance and policy example into the questions a board should put to management. Use it to assess completeness: any row without a confident, evidenced answer marks a gap in principles, policy or oversight that the board should direct management to close.
| Dimension | What good looks like | Board's critical question | Qualification to respect |
|---|---|---|---|
| Fairness (Ethical Principles) | AI systems designed and used for equitable, unbiased outcomes, actively avoiding discrimination in HR processes | How do we detect and mitigate bias in the data and algorithms behind our recruitment and performance tools? | Requires continuous monitoring and mitigation of potential biases in data and algorithms |
| Transparency (Ethical Principles) | AI operation and decision-making understandable and explainable to relevant stakeholders, especially in HR applications | Do candidates and employees know when AI influenced a decision, and can we explain the rationale? | Degree of transparency may vary with system complexity and proprietary considerations, but the rationale should be clear |
| Accountability (Ethical Principles) | Clear lines of responsibility for AI development, deployment and outcomes in the HR function | Who is named as responsible when an AI-assisted HR decision goes wrong? | Accountability extends to adherence to ethical pillars and monitoring AI usage for compliance |
| Policy scope and compliance (Policy Components) | Policy covers in-house, procured and embedded AI; aligns to frameworks such as the EU AI Act, NIST AI RMF and OECD AI Principles | Does our policy cover AI embedded in vendor HR tools, not just systems we build? | Alignment to named frameworks is a policy statement, not proof of compliance with any specific regulation |
| Risk assessment and audit (Policy Components) | Defined cadence for AI risk assessments and audits, with third-party audits for high-risk systems | What is our audit cadence for HR AI, and when do external audits trigger? | High-risk HR use cases warrant stricter audit treatment than low-risk ones |
| Cross-functional committee (Oversight Structure) | A dedicated AI governing body establishing accountability and monitoring AI strategy within HR | Does our committee include risk, legal, IT, security, HR and business representation — and does it have decision rights over new use cases and training? | A committee without decision rights is the identified pitfall |
| Board-level reporting (Oversight Structure) | Continuous monitoring and regular board reporting on responsible, ethical AI adoption and compliance | What do we receive each cycle on adherence to principles, usage audits and anomalies? | Reporting should cover adherence to ethical pillars, usage audits and identification of anomalies |
Read the matrix as a gap diagnostic rather than a checklist to file away. A board that can answer the fairness and committee rows but not the policy-scope row has a vendor-management blind spot; a board with a policy but no committee decision rights has governance on paper only. The framework is deliberately vendor-neutral and jurisdiction-neutral: it describes governance structure, not technical implementation, and it is not legal advice on compliance with any specific regulation such as the EU AI Act.
Actionable Steps for Board Leadership
Boards should initiate or review their organization's AI governance framework for HR, ensuring it formally addresses ethical principles, comprehensive policies, and robust oversight mechanisms, including a dedicated cross-functional committee. The measurable indicator of success is concrete: the establishment of a documented AI governance policy for HR, a formally constituted AI ethics and compliance committee with defined responsibilities, and regular board reporting on AI risks and compliance. When all three exist and the committee holds authority over new-use-case review, the board can credibly claim to be governing HR AI rather than merely observing it — and the next board cycle's report on usage audits and anomalies becomes the evidence that the system is working.





