The AI Literacy Illusion: Why Board Education Isn’t the Problem
- Mar 16
- 7 min read
Week 4 demonstrated risk differentiation: different AI systems yield fundamentally different consequences, requiring distinct governance approaches. Massive reader response raised another question: What capability do boards actually need?
The conventional answer: AI literacy. Send directors to courses. Build technical understanding.
But research reveals something uncomfortable: the organisations governing AI most effectively aren’t those with the most AI-literate boards.
The Wrong Job Advertisement
Consider two board recruitment advertisements:
Advertisement A: “Seeking board member with PhD in machine learning, 10+ years AI industry experience, deep technical knowledge of large language models…”
Advertisement B: “Seeking director who can challenge management’s AI assumptions, demand evidence of value creation, recognise governance gaps, and apply fiduciary duties to intelligent systems.”
McKinsey finds 66% of directors have limited to no AI knowledge. Yet 39% of Fortune 100 companies successfully disclose board oversight, even though many directors possess minimal AI expertise.
The correlation everyone assumes – between director AI knowledge and governance effectiveness – research doesn’t demonstrate.
Perhaps we’re solving the wrong problem.
How We Got Here
Organisations invest millions in AI literacy programmes. After significant investment, McKinsey reports 66% of directors still report limited knowledge. Only 15% of boards receive meaningful AI metrics. Less than 25% have structured policies with scaling rules.
Education investment isn’t translating to governance capability. Why? AI evolves faster than courses can teach. Technical knowledge doesn’t translate to governance judgment. Directors becoming “AI literate” doesn’t make management governable. Focus on knowledge acquisition distracts from building systems that actually work.
Dr Rumman Chowdhury: “Directors don’t need to be AI experts but must understand implications.”
Governance expertise matters more than technical knowledge.
What Boards Already Do
Boards already govern complex matters without technical expertise:
Pharmaceutical companies: Boards without chemistry PhDs ensure clinical trials function, regulatory compliance holds, and safety monitoring works. They demand evidence of efficacy and risk assessments. They apply governance expertise to scientific complexity.
Financial services: Boards without quantitative backgrounds govern derivatives trading. They ensure risk frameworks operate correctly and demand that value-at-risk be translated into business impact. They apply governance expertise to mathematical complexity.
Healthcare: Boards without medical degrees govern clinical services. They ensure quality assurance works and that patient safety receives priority. They demand outcome data and incident reviews. They apply governance expertise to clinical complexity.
Australian Institute of Company Directors: duty of care requires reasonable inquiry and judgment, not technical expertise.
Directors can’t hide behind “I don’t understand AI.” But they needn’t become AI experts either.
The Four Real Competencies
Strategic Fluency – Not Technical Literacy
Effective boards understand business value propositions, not algorithmic mechanics. They recognise when AI creates value rather than destroys it. They spot governance gaps and connect AI strategy to organisational strategy.
Financial services boards don’t need to understand fraud algorithms. They need to understand whether deployment reduces losses, at what cost, with what impact on customers, and what happens when systems err.
Governance Craft in Dynamic Environments
AI requires faster governance cycles. Boards must govern systems that learn and evolve. Make decisions with incomplete information. Oversee rapid deployment cycles measured in weeks, not months.
These represent skill development in governance, not AI education requirements.
The Right Questions
Effective boards ask second-order questions:
Not: “How does this AI system work?” (technical)
But: “How do we know it’s working as intended? What’s our evidence?” (governance)
Not: “What’s the accuracy?” (metric)
But: “What decisions rest on this 94% accuracy? What’s our protocol for the 6% errors? Do errors distribute randomly or discriminate systematically?” (implications)
Not: “Can AI do X?” (capability)
But: “Should AI do X for our organisation? Who decides? What’s our risk appetite?” (accountability)
Board proposes AI for hiring. Director asks: “Show me evidence this doesn’t discriminate. What’s your testing methodology? Who validated it? What if bias is discovered post-deployment? Who’s accountable for legal exposure?”
Governance questions require governance expertise, not AI expertise.
Demanding Better Intelligence
The ready 14% from Week 1 don’t have more AI knowledge. They have better governance intelligence: usage metrics showing what’s happening, outcome tracking showing value created, risk indicators flagging problems, stakeholder feedback demonstrating impact.
BCG: “AI fluency isn’t about turning directors into technologists – it’s ensuring they can scrutinise management’s logic with confidence.”
Building Governance Infrastructure
Successful boards invest in better reporting frameworks (Joe Knight’s KPI dashboards), clearer escalation protocols (McKinsey’s triggers), robust inventory systems (IOD continuous audits), outcome measurement (BCG dashboards) and usage monitoring (Scott Bridgen’s real metrics).
Not AI technical courses. Not machine learning boot camps. Not algorithmic mathematics.
The capability gap is in governance infrastructure, not the knowledge base.
Rajjie Sarmey: “Boards need documented proof of governance in action, not technical credentials.”
Implications for Board Development
Stop sending directors to generic AI courses. Start building governance infrastructure that makes AI governable.
Instead of technical education, invest in scenario-planning exercises that present governance dilemmas. “AI system creates discriminatory outcomes post-deployment. Walk through response.”
Invest in peer learning to show how other boards govern complexity they don’t understand. Healthcare boards governing clinical services. Financial boards governing derivatives.
Invest in process design workshops to build escalation protocols, scaling rules, and monitoring systems. Hands-on governance infrastructure design.
Invest in strategic gaming testing and governance to respond to AI incidents. Simulated crises reveal governance gaps.
McKinsey emphasises benchmark relentlessly. Learn from boards governing complex matters well, regardless of sector. Governance craft transfers. Technical knowledge doesn’t.
The Choice
The AI literacy obsession solves the wrong problem.
Boards don’t need to understand how AI works. They need to govern organisations where AI works.
66% of directors lack AI knowledge. That might be acceptable if we stop confusing knowledge gaps with governance gaps.
Week 4 demonstrated that different AI risks need different governance. Week 5 reveals why boards can implement risk-tiered governance without AI expertise: they already possess governance expertise developed throughout their careers.
The pharmaceutical director, without a PhD in chemistry, ensures clinical trials function. The financial director, without a quant background, ensures risk management works. The healthcare director without a medical degree who ensures patient safety receives priority.
They apply governance expertise to technical domains. AI requires the same approach.
The real solution: Build governance infrastructure. Demand better intelligence. Apply existing governance expertise to a new context.
Next week: the counter-intuitive finding that strong governance doesn’t constrain AI adoption – it accelerates it. The ready 14% discovered governance as a growth engine, not a bureaucratic burden.
Will you invest another year building AI literacy whilst governance gaps widen? Or will you build governance infrastructure that makes AI governable now?
The evidence is clear. The path is defined. The choice is yours.
Reflection Questions for Your Board
Infrastructure vs Education: Has your organisation invested more in teaching directors about AI, or in building governance infrastructure (comprehensive inventories, real-time metrics, clear escalation protocols) that makes AI governable?
Governance Capability: Can your board identify specific governance gaps in AI deployment proposals, or do directors feel unable to challenge management without more technical knowledge?
Board Time Allocation: What percentage of board time discussing AI focuses on how systems work versus whether governance processes are functioning effectively?
Evidence vs Assurance: When was the last time your board received evidence, not assurance, that AI governance is actually working (usage data, incident rates, bias testing results, stakeholder impact)?
Cross-Sector Comparison: If your pharmaceutical/financial/healthcare governance succeeds without directors possessing technical degrees in chemistry/quantitative finance/medicine, why does AI governance require directors to become AI experts?
About the Series
This is Week 5 of our AI Governance series examining how boards can move from governance structures to governance readiness.
Article 4 When AI Errors Cost Lives, Customers, Or Markets: The Board’s Guide to Risk-Proportionate Governance
Article 5 questions the AI literacy Illusion.
Next article: Exploring how governance accelerates AI deployment – the ready 14% discovered governance as a competitive advantage, not a constraint.
About the Author
Viren Lall is Managing Director of ChangeSchool LDN, an EFMD award-winning executive education delivery partner. ChangeSchool develops transformational AI capability for leaders and boards through discovery-based approaches that bridge academic rigour with operational reality.
ChangeSchool’s research into AI governance best practices informs our work with boards and executive teams across engineering, manufacturing, and education sectors. Through discovery-based approaches, we help organisations move from governance structures to governance readiness. If your board is navigating AI governance challenges, I’d welcome the opportunity to discuss them.
Acknowledgments
Insights from this article draw on research and expertise from:
Dr Rumman Chowdhury – Directors need to understand implications, not become AI experts
Boston Consulting Group (BCG) – AI fluency as scrutinising management logic with confidence, not technical knowledge acquisition
Rajjie Sarmey – Boards need documented proof of governance in action, not technical credentials
Australian Institute of Company Directors (IoCD) – Duty of care requires reasonable inquiry and judgment, not technical expertise.
Joe Knight, FTI Consulting – KPI dashboards for measuring governance effectiveness, not just governance existence
Scott Bridgen – Real usage monitoring metrics revealing operational reality
Sources
McKinsey & Company (2024). “The State of AI in 2024: Generative AI’s Breakout Year.” McKinsey Global Survey. Available at: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
Sedgwick (2025). “2026 Global Risk Report: Forecasting Report.” Survey of 300 senior leaders at Fortune 500 companies.
Boston Consulting Group (2025). “From Observer to Multiplier: AI Governance for Boards.” BCG Analysis and Insights. Available at: https://www.bcg.com
Australian Institute of Company Directors (2025). “AI Use by Directors and Boards.” IoCD Research Report. Available at: https://www.aicd.com.au
Knight, Joe (2024). “AI Governance KPIs: Measuring What Matters.” FTI Consulting Insights.
Institute of Directors UK (2024). “AI Governance in the Boardroom.” IoD Research Report. Available at: https://www.iod.com
