**THE EDITORIAL BOARD**

*Board-Level AI Governance: Independent Advisory for MENA Corporate Directors*

Corporate directors across the Middle East and North Africa are being asked to approve AI strategies they do not fully understand, overseeing risks they cannot measure, and signing off on investments whose failure modes are invisible to traditional audit committees. The gap between board responsibility and board competence in artificial-intelligence governance has become the single most consequential governance deficiency in the region’s largest enterprises. Independent advisory structures—external specialists embedded at board level—are emerging as the mechanism through which MENA boards close that gap. The design of those structures determines whether AI governance becomes a genuine control system or a compliance checkbox.

The board-level AI governance advisory gap is not a knowledge deficit that can be solved through occasional executive education. It is a structural mismatch between the velocity of AI deployment and the tempo of board oversight. MENA corporate directors, drawn largely from finance, energy, and construction backgrounds, are now governing organisations that deploy machine-learning models across customer credit, supply-chain optimisation, predictive maintenance, workforce planning, and regulatory compliance. The technical depth required to interrogate those models, the ethical sensitivity required to question their societal impact, and the strategic foresight required to align AI investment with long-term enterprise value are not skills acquired in a two-day governance seminar. They are the product of sustained, independent advisory relationships.

Data from McKinsey supports the urgency. In its 2024 Global Board Survey, 67 percent of directors in the GCC and 58 percent in North Africa identified AI governance as the top oversight challenge they feel least prepared to address. Gartner’s 2025 CIO Survey found that 82 percent of large MENA enterprises have deployed AI in at least one business unit, but only 23 percent have board-level AI literacy programmes, and fewer than 12 percent have dedicated external advisory arrangements for AI oversight. PwC’s Middle East Governance Outlook 2025 notes that independent AI advisers are present at board level in fewer than 8 percent of publicly listed MENA companies, compared with 31 percent in Western Europe and 27 percent in North America. The underrepresentation is not a reflection of lower AI adoption; it is a reflection of oversight structures that have not kept pace with deployment reality.

The consequences of governance-by-wishful-thinking are material. BCG reports that AI-related regulatory penalties in the financial services, healthcare, and telecommunications sectors across the GCC, Saudi Arabia, and Egypt have risen threefold since 2022. Board liability exposure around AI decision-making is increasing as regulators treat algorithmic bias, data-privacy violations, and autonomous system failures as governance failures rather than operational errors. Directors who cannot demonstrate informed oversight of AI systems are finding themselves personally exposed in ways that traditional D&O insurance policies are only beginning to address.

—
Governance advisory gap

The governance advisory gap manifests in three distinct dimensions: knowledge asymmetry, temporal misalignment, and structural absence.

Knowledge asymmetry originates in the technical complexity of modern AI systems. A board reviewing a credit-scoring model, a predictive-maintenance deployment, or an AI-assisted hiring tool must understand not only how the model arrives at its outputs but also the statistical properties of its training data, the nature of its feature engineering, the conditions under which it fails, and the institutional incentives of the vendors who built it. Directors with strong financial literacy can read a balance sheet and identify earnings manipulation because decades of accounting standards have made that knowledge portable and comparable. No equivalent framework exists for AI model inspection. The result is that boards approve AI budgets and strategies with less scrutiny than they apply to a ten-million-dollar construction contract.

Temporal misalignment arises because AI deployment cycles are dramatically shorter than the reporting and review cycles through which boards traditionally exercise oversight. A MENA enterprise might deploy an AI customer-service chatbot in four weeks, retrain its demand-forecasting model every quarter, and update its fraud-detection algorithm monthly. Board meetings occur quarterly, audit committee reviews are semiannual, and strategic planning cycles are annual. The governance architecture assumes slower, more stable systems. When systems evolve faster than oversight, governance becomes retrospective rather than prospective: by the time the board learns of a model failure or a data incident, the damage is embedded in customer relationships, regulatory filings, or operational workflows. Proactive governance of AI requires continuous monitoring and rapid escalation mechanisms that traditional board structures do not provide.

Structural absence is the most fundamental gap. MENA boards typically rely on three standing committees—audit, risk, and nomination and remuneration—to deploy specialist expertise. None of those committees is designed for AI oversight. The audit committee validates financial controls. The risk committee reviews enterprise risk through a lens shaped by insurance, compliance, and operational continuity frameworks. The nomination and remuneration committee evaluates director and executive competence through the filters of industry experience, financial performance, and leadership track record. AI governance spans all three: it has financial implications, operational risks, and talent requirements. Yet it belongs to none. This structural gap means that AI oversight is either inadequately distributed across committees with competing priorities or omitted entirely, left to informal CEO-board conversations that fail to satisfy governance codes or regulatory expectations.

—
Why boards need independent AI oversight

The case for independent AI oversight rests on four pillars: model risk, regulatory exposure, stakeholder trust, and strategic coherence.

Model risk is the most immediate concern. AI models fail in ways that are interpretable only to those who understand their construction. A consumer-credit model trained on historical lending data that encodes historical bias against female applicants or expatriate workers can systematically underwrite profitable customers while violating fair-lending principles. A predictive-maintenance model trained on temperate-climate equipment data may miss failure modes specific to high-temperature, dusty MENA operational environments. A supply-chain optimisation model that learns from pandemic-era logistics data may embed volatile demand assumptions that destroy value when conditions normalise. Without independent examination at board level, these failure modes remain invisible until they generate material losses.

Regulatory exposure is accelerating across MENA markets. Saudi Arabia’s PDPL imposes strict obligations on automated decision-making systems that process personal data. The UAE’s AI Ethics Principles, issued by the UAE AI Office, require ethical impact assessments for high-risk AI deployments. Egypt’s draft Data Protection Law includes provisions specific to algorithmic profiling. Qatar’s National AI Strategy calls for mandatory AI governance reviews in government-linked entities. The trend is globally consistent: regulators are moving from voluntary guidelines to binding obligations, and board accountability is being explicitly extended to AI governance failures. Independent advisers provide the technical bridge between board-level fiduciary duty and the granular AI compliance requirements that general counsel and external auditors are not equipped to satisfy.

Stakeholder trust is the strategic asset at risk. MENA enterprises operate in markets where sovereign ownership, paternalistic governance traditions, and family-business reputational concerns create high stakes for social legitimacy. A UAE bank deploying AI-driven customer segmentation that appears to discriminate by nationality faces reputational damage that quickly becomes political. A Saudi retailer using AI demand forecasting that systematically underinvests in rural regions provokes regulatory scrutiny and media criticism. A Moroccan manufacturer whose AI quality-control system falsely rejects domestic-product variants while accepting exports triggers consumer boycotts and parliamentary questions. The common thread is that AI failures, when they occur, are socio-technical events: they interact with discrimination norms, regional equity expectations, and brand relationships that extend beyond pure financial risk. Independent advisers provide the nuanced contextual understanding required to anticipate these stakeholder dynamics.

Strategic coherence distinguishes enterprises that treat AI as a technology investment from those that treat it as a core strategic capability. Board-level AI advisory that is embedded into governance processes—rather than consulted ad hoc—forces alignment between AI deployment and enterprise strategy. Advisers who attend board presentations, review model validation plans, and participate in investment approvals ask questions that operational AI teams and technology vendors cannot ask themselves: Does this model create competitive advantage that other firms cannot replicate? Does it centralise or distribute strategic control? Does it concentrate risk in ways that the enterprise’s capital structure does not support? Without that external perspective, AI strategy defaults to incremental adoption driven by operational pressure rather than deliberate design driven by board-level intent.

—
Advisory mandate design

The design of an independent AI governance advisory mandate determines its practical effectiveness. Generic advisory arrangements—an external consultant presenting an annual slide deck—do not constitute governance. Effective mandates must specify access, scope, authority, and reporting.

Access must be both timely and candid. Advisers should be entitled to board-level information before board decisions are finalised, not after. This means receiving model documentation, risk assessments, and regulatory filings before they are presented to the board, and being able to provide observations directly to the chair or to a committee. The access right must be codified in the board charter and reinforced by the chair’s institutional support. An independent adviser who learns of a major AI model deployment after the board has approved it is performing audit, not governance.

Scope must cover the full AI lifecycle: strategy, design, deployment, monitoring, and decommissioning. Narrow scopes that limit advisory review to post-deployment compliance checks miss the strategic and ethical dimensions that are most valuable. In MENA, scope should explicitly include regulatory mapping across jurisdictions where the enterprise operates, allowing advisers to flag cross-border data-transfer risks, multi-jurisdictional ethical standards, and regional sovereign AI requirements. Scope must also include vendor oversight, because many MENA enterprises rely on international AI vendors whose models are trained on data and assumptions that may not align with local operating environments.

Authority is the most difficult mandate element because boards, rightly protective of their own decision-making primacy, often resist granting external advisers formal approval or veto rights. The workable solution is graduated authority: advisers submit advisory reports to the board or committee, the board considers those reports within a defined timeframe—typically two to four weeks—and the board must either approve, reject, or formally note the advice with documented rationale. This creates governance accountability without transferring fiduciary responsibility. Documentation of the board’s response to independent advice is itself a governance record that regulators and auditors increasingly require.

Reporting cadence must reflect AI deployment velocity. Quarterly reporting aligns with audit committee cycles for high-volume, low-risk models; monthly reporting is appropriate for medium-risk operational AI; and ad hoc reporting on an alert-driven basis is necessary for high-risk models affecting customer welfare, regulatory compliance, or strategic decisions. The reporting framework must include not only advisory observations but also a board response log, creating an auditable trail of how independent advice influenced decisions. BCG’s 2023 MENA AI Governance Benchmark found that enterprises with documented board responses to independent AI advice had 40 percent fewer AI-related regulatory incidents than enterprises with advisory functions that lacked formal response requirements.

—
Committee structure

The structural absence identified earlier requires committee design that translates AI governance obligations into the MENA board template. Three structural options merit consideration: a dedicated AI governance committee, an augmented audit committee, and a joint risk-technology committee.

A dedicated AI governance committee is the cleanest structural solution. It centralises oversight responsibility, provides a home for the independent adviser, and signals to stakeholders—regulators, investors, employees—that AI governance is a board-level priority. The committee typically comprises three to five directors, at least one of whom has demonstrable AI or digital-technology competence. The independent adviser attends all meetings as a standing participant but does not vote. The committee reviews AI strategy, approves high-risk model deployments, monitors regulatory compliance, and reports to the full board. In MENA, where board committees must balance national representation requirements with capability needs, a dedicated committee allows enterprises to combine directors with relevant sector expertise with the independent adviser’s specialist knowledge without forcing either into an unfamiliar committee context.

An augmented audit committee is the most pragmatic short-term option for enterprises that lack the scale or governance maturity for a new committee. The audit committee adds AI oversight to its existing mandate, expanding its charter to cover model risk, algorithmic compliance, and AI-related financial reporting. The independent adviser supports the audit committee rather than sitting in a separate forum. The limitation is workload: audit committees in MENA often carry heavy financial reporting and regulatory compliance burdens, and adding AI oversight without corresponding resourcing degrades the quality of all oversight. Enterprises that choose augmentation should assign at least one audit committee member to AI oversight as a designated focus area, creating accountability within the existing structure.

A joint risk-technology committee combines risk-committee governance discipline with technology-committee technical literacy. It is particularly suitable for enterprises with substantial technology portfolios spanning cybersecurity, digital infrastructure, and AI. The independent adviser co-chairs or leads technical discussions within the committee. The advantage is integration: AI governance is reviewed alongside enterprise risk management rather than in isolation. The risk is scope creep; risk-technology committees can become discussion forums rather than decision bodies if not carefully managed with clear terms of reference and meeting protocols.

Gartner’s recommendation for MENA boards is a dedicated committee within three years for enterprises with AI investments above ten percent of capital expenditure, and an augmented audit committee for enterprises below that threshold. The threshold is calibrated to the point at which AI risk becomes material to enterprise valuation. For the region’s large family-owned conglomerates, sovereign-linked enterprises, and financial institutions, the threshold is often already crossed.

—
Information flows

Governance is only as effective as the information it receives. Board-level AI advisory depends on information flows that are technical, contextual, timely, and structured in ways that busy directors can process.

Technical information must move upward from implementation teams to the board in summaries that respect directors’ time while preserving decision-relevant detail. The standard reporting format—a project status update with a green, amber, or red indicator—is inadequate for AI governance. Directors need information about training-data provenance, model-validation results, performance degradation curves, regulatory impact assessments, and vendor accountability clauses. The independent adviser is best positioned to translate this technical material into board-comprehensible formats that highlight decision points and risk concentrations rather than overwhelming directors with implementation minutiae.

Contextual information bridges the gap between model behaviour and enterprise-specific operating conditions. An AI model validated against benchmark datasets may perform differently in Cairo traffic, a Jeddah hospital, or a Casablanca factory because local data distributions differ from global training sets. Contextual briefing materials prepared by the independent adviser should include local-environment performance summaries, regional regulatory requirement matrices, and stakeholder impact assessments that explain how a model’s outputs affect customers, employees, regulators, and communities in MENA operating contexts. Without this contextual layer, boards receive technically precise information that is strategically misleading.

Timeliness means that AI incidents—model failures, data-quality events, regulatory inquiries—are escalated to the independent adviser within defined SLA windows, typically four hours for critical incidents and twenty-four hours for significant incidents. The adviser then determines whether the incident warrants board or committee notification. Escalation protocols should be codified in the AI incident response plan and should specify the decision criteria, notification thresholds, and communication formats. Timeliness is particularly important in MENA because regulatory reporting deadlines for data incidents are often compressed compared with European frameworks, and media exposure of AI-related fairness or privacy issues spreads rapidly across social platforms.

Structured formats reduce cognitive load and decision errors. The MENA Enterprise AI Governance Information Standard, proposed by the Dubai International Financial Centre’s AI Regulator and adopted in pilot form by several Riyadh-based institutions, defines a five-section mandatory briefing format: strategic alignment, risk summary, regulatory status, stakeholder impact, and board decision required. Advisers should use this format for all committee and board submissions. Consistency in format allows directors to process AI information faster, compare across quarters, and identify trends that would otherwise be obscured by stylistic variation.

—
Risk and compliance

AI risk governance in MENA must address the intersections of operational risk, data-privacy risk, regulatory compliance, and reputational risk in ways that are specific to the region’s regulatory mosaic and market dynamics.

Regulatory compliance requires explicit mapping of AI deployments against applicable laws and emerging regulations. Beyond PDPL, the UAE’s AI Ethics Principles, and Egypt’s data-protection provisions, enterprises must track European and North American regulations that affect MENA operations: the EU AI Act applies to any enterprise selling products or services into the EU, including Saudi Aramco’s commercial subsidiaries, Emirates Airlines, and Morocco’s automotive exporters. The EU AI Act’s risk-tiered framework—prohibited, high-risk, limited-risk, and minimal-risk categories—requires AI model classification, conformity assessments, and registration obligations for high-risk systems. MENA enterprises with EU market exposure need independent advisory support to classify their AI models correctly and to design the compliance documentation that regulators increasingly require. Cross-border data-transfer governance is a related compliance imperative: PDPL, UAE data-protection law, and Egypt’s draft regulations all impose restrictions on personal-data flows that constrain AI training pipelines and model-inference architectures.

Operational risk centres on model reliability in local operating environments. AI models that perform well in benchmark testing may degrade when deployed on operational data with distribution shifts caused by seasonal variations, economic fluctuations, or demographic changes specific to MENA markets. Model-monitoring frameworks must establish performance thresholds, define retraining triggers, and ensure that degradation events are detected before they generate customer harm or regulatory exposure. Independent advisers should require model-validation evidence that includes local-environment testing, not just global benchmark results.

Data-privacy risk is multifaceted. Beyond the legal obligations, MENA enterprises face cultural expectations around personal-data stewardship that exceed many regulatory minimums. Corporate leaders in the region understand that data misuse erodes the social licence on which large enterprises—particularly those with partial state ownership—depend. Advisers should treat privacy governance as a strategic rather than merely compliance exercise, ensuring that data-governance frameworks support both regulatory requirements and stakeholder trust.

Reputational risk is amplified in MENA by concentrated media ownership, social-media amplification, and the political sensitivity of algorithmic decisions that affect employment, credit access, or public-service distribution. An independent adviser should conduct pre-deployment reputational risk assessments for high-profile AI deployments, identifying likely stakeholder concerns and recommending mitigation strategies before incidents occur. BCG’s research shows that organisations that conduct pre-deployment stakeholder impact assessments reduce AI-related media incidents by 60 percent compared with organisations that react to controversies after launch.

—
Monitoring and reporting

Continuous monitoring and structured reporting are the operational mechanisms through which board-level AI advisory translates into governance outcomes. They must be calibrated to risk level, built on objective indicators, and supported by independent verification.

Monitoring frameworks should adopt the three-line model familiar to MENA audit committees. The first line—operational teams—manages day-to-day model performance and data-quality checks. The second line—risk, compliance, and data-governance functions—reviews first-line reports, validates monitoring methodologies, and escalates anomalies. The third line—internal audit or the independent adviser—provides independent assurance that the first and second lines are operating effectively. For AI governance, this three-line model translates into model-performance dashboards reviewed weekly by operational teams, monthly by risk and compliance functions, and quarterly by the independent adviser and board committee. The independent adviser’s quarterly review should include not only performance metrics but also an assessment of the integrity of the monitoring process itself: Are the right models being monitored? Are the right thresholds being used? Is the process resistant to gaming by teams whose performance is being measured?

Reporting to the board must follow the cadence and format requirements specified in the information-flows section. Advisory reports should include: executive summary, risk rating, compliance status against applicable regulations, performance trends compared with previous periods, emerging risks, and recommended actions. Advisers should avoid presenting raw technical data; instead, they should translate technical findings into board-actionable insights. A model whose fraud-detection accuracy has fallen by four percentage points requires explanation of the financial exposure, the likely causes, and the remediation options—not the ROC curve itself.

Escalation thresholds create decision rights during incidents. A common framework, recommended by McKinsey for MENA enterprises, defines three thresholds. Yellow escalation occurs when a model’s performance falls within a pre-specified warning band: a five-percentage-point drop in accuracy, a ten-percent increase in false-positive rate, or a data-quality score below ninety-five percent. Yellow events trigger adviser investigation and committee briefing within the next reporting cycle. Red escalation occurs when a model breaches performance thresholds that constitute regulatory non-compliance or material financial exposure: a twenty-percentage-point accuracy drop, a regulatory inquiry, or a customer-harm event. Red events require immediate adviser notification, emergency committee briefing, and board decision within forty-eight hours. Black escalation—an event that threatens enterprise continuity, such as a systemic AI failure across multiple business lines—triggers full board convening authority. Defining these thresholds in advance removes discretion from operational teams under pressure and creates predictable governance behaviour.

—
Adviser selection

The effectiveness of board-level AI governance depends substantially on the quality of the independent adviser. Selection criteria must address technical competence, regulatory knowledge, MENA contextual understanding, independence, and relational fit.

Technical competence requires demonstrated expertise in AI model development, validation, and risk management—not solely in AI strategy consulting or vendor implementation. The ideal adviser can read model documentation, assess validation methodology, and identify failure modes that operational teams and auditors miss. This depth of technical skill must be supplemented by system-level thinking: the ability to understand how multiple models interact across an enterprise value chain, creating compound risks that individual model reviews would not capture.

Regulatory knowledge must span the full set of jurisdictions relevant to the enterprise. For a Saudi conglomerate with operations across the GCC, Egypt, and Europe, the adviser must understand PDPL, UAE AI Ethics, EU AI Act requirements, and applicable Egyptian regulations. Regulatory requirements for AI governance are evolving rapidly across all these jurisdictions; advisers must be actively engaged in regulatory development rather than relying on static compliance memoranda.

MENA contextual understanding is a differentiation factor that separates effective local advisers from global generalist firms. An adviser who understands the region’s regulatory culture—its combination of rapid legislative innovation, discretionary enforcement, and relationship-based accountability—can navigate compliance requirements more effectively than one who approaches MENA markets through a purely Western regulatory lens. Contextual understanding also encompasses operating environments: the demographic, linguistic, and economic conditions that affect AI model behaviour in MENA markets.

Independence is non-negotiable but frequently compromised. Advisers who also consult to the enterprise’s technology vendors, who maintain financial relationships with AI vendors that the enterprise is procuring, or who are selected by the CEO rather than the board itself cannot provide the objective perspective that board-level governance requires. Selection processes should be owned by the board committee responsible for AI governance, not by management. Contracts should include conflict-of-interest certifications, rotation clauses, and explicit protections against vendor influence. Independent advisers should be selected from firms whose primary service line is governance and risk, not implementation consulting, to avoid the structural conflict that arises when an adviser audits its own advice.

Relational fit matters because the adviser must work closely with the board chair, committee chairs, and general counsel. The selection process should include interactions with these key relationships before appointment. Advisers should be evaluated on their ability to explain complex technical material to non-technical audiences, to challenge management assertions without destroying trust, and to maintain perspective under political or commercial pressure.

McKinsey’s 2023 MENA AI Governance Practice Study found that enterprises that selected independent advisers through competitive procurement processes managed by the board committee rather than by the CEO had 35 percent fewer AI governance failures over a three-year observation period than enterprises where management selected the adviser unilaterally. The difference reflects not adviser quality but structural independence: when advisers are accountable to the board rather than to management, their reports are more candid and their advice more consequential.

—
MENA board governance roadmap

A practical roadmap for implementing board-level AI governance in MENA enterprises spans three years and aligns with enterprise AI maturity evolution. It has four phases: Awareness, Structure, Operating, and Embedded.

Phase 1: Awareness, months one to four. Activities include a board-level AI literacy programme facilitated by the independent adviser, an AI asset inventory mapping all models deployed across business units, a regulatory gap analysis identifying applicable MENA and international AI governance requirements, and a preliminary risk assessment ranking models by enterprise impact. Deliverables include a board AI literacy certificate, an AI asset register, a regulatory compliance matrix, and a risk-rating schedule. The phase cost is moderate and can often be funded from existing learning and development budgets.

Phase 2: Structure, months five to twelve. Activities include committee design and charter amendment, selection and appointment of the independent adviser, development of AI governance policies covering model approval, monitoring, and incident escalation, and ratification of information-flow protocols specifying what AI information reaches the board, in what format, and on what cadence. This phase requires board-level decisions about committee structure, adviser selection methodology, and reporting requirements. The output is a formalised AI governance framework that can be communicated to regulators, investors, and stakeholders.

Phase 3: Operating, months thirteen to twenty-four. The committee and adviser move from planning to active oversight: reviewing model deployments, monitoring regulatory compliance, validating model validation methodologies, and reporting to the board. Operating discipline includes maintaining model registers, conducting quarterly model audits, updating regulatory requirement matrices, and managing adviser performance. In this phase, enterprises should develop model-risk appetite statements that define acceptable performance thresholds and regulatory exposure levels, giving the committee and adviser clear decision boundaries. PwC’s MENA governance maturity model shows that enterprises that achieve operating-phase rigour within twenty-four months reduce AI-related regulatory incidents by more than 50 percent compared with enterprises that take longer to establish committee processes.

Phase 4: Embedded, months twenty-five to thirty-six. By this phase, AI governance should be indistinguishable from enterprise governance. Rather than treating AI as a technology governance issue embedded in a dedicated committee, governance frameworks should integrate AI oversight into the risk, audit, and strategic-planning cycles of the full board and all standing committees. The independent adviser transitions from a standing committee participant to a specialist resource accessed across all governance activities. At this stage, the adviser also supports the nomination and remuneration committee by informing director-competency requirements and executive-performance evaluation criteria that include AI governance capability.

Regional variations affect roadmap implementation. Gulf Cooperation Council enterprises—particularly those with sovereign links, financial services exposure, or EU market operations—can move rapidly because regulatory pressure and stakeholder expectations are most developed in those markets. Saudi enterprises, driven by Vision 2030 digital transformation requirements and the scale of sovereign-linked AI investment, are positioned to lead regional adoption. North African enterprises face less immediate regulatory pressure but should establish foundations now because EU export market exposure and local digitisation programmes will accelerate governance requirements over the roadmap horizon. Enterprises that delay until regulatory requirements crystallise will spend more on compliance remediation than those that build governance proactively.

The roadmap is an evolution, not a project with a defined end-state. AI technology, regulatory requirements, and enterprise AI portfolios will all continue to evolve. Board-level AI governance is a permanent organisational capability, not a one-time implementation. The enterprises that treat it as a strategic investment in governance infrastructure—rather than a compliance cost—will find that it strengthens board effectiveness, reduces regulatory exposure, and builds the stakeholder trust that determines long-term enterprise value in MENA markets increasingly sensitive to algorithmic fairness, data stewardship, and responsible technology deployment.

Table: Board-Level AI Governance Advisory Mapped to Board Committee and Responsibility
Committee: Dedicated AI Governance Committee / Audit Committee (Augmented) / Risk Committee / Nomination and Remuneration Committee
Responsibility: Model strategy and deployment approval / Data privacy compliance and financial reporting / Enterprise AI risk framework and regulatory mapping / AI governance competency in director recruitment and executive evaluation

—