# Board AI Literacy Curriculum and Self-Assessment Templates Building enough AI literacy at board level for real oversight. A curriculum outline tied to the company's own AI risks, an annual board briefing template that is context, not operations, a director self-assessment, and criteria for choosing an external educator. **Who this is for:** The company secretary, general counsel, or chair building the board's AI oversight capacity. Source playbook: https://aigovernance.com/playbook/director-ai-literacy --- ## Director AI literacy curriculum outline _Topics, depth, and how each connects to the specific AI risks this company runs._ ### Template | Topic | Depth for directors | Connection to our AI risks | |---|---|---| | What AI can and cannot do (capabilities and limits) | conceptual | which of our decisions rely on it | | How AI systems fail (bias, drift, hallucination, brittleness) | conceptual + examples | our people-affecting and customer-facing systems | | The regulatory landscape | working knowledge of the regimes we are in | EU AI Act, US state law, sector rules for our sectors | | Governance and oversight structures | working knowledge | our committee, three-lines model, reporting | | Reading an AI risk report | practical | our board dashboard and thresholds | | Escalation and crisis | practical | our escalation register and incident types | ### Worked example > Tailoring for our board (AI in hiring and fraud decisions, both people-affecting and regulated). | Topic | How we tailored it | |---|---| | How AI fails | leads with a bias case study built from our own adverse-impact history, anonymised | | Regulatory landscape | covers only the EU AI Act, Colorado SB205, NYC LL144, and fair-lending analogues, not a global survey | | Reading an AI risk report | uses our actual board dashboard and thresholds | ### Acceptance criteria - Every topic states the depth expected of a director, not a practitioner. - Each topic is connected to a specific AI risk the company actually carries. - The curriculum uses the company's own systems and reports as teaching material. --- ## Board AI briefing template _An annual context-setting session, deliberately separate from the quarterly operational report._ ### Template > Once a year, 60-90 minutes. Sets context, not status. - **Where AI sits in the strategy:** what the company is betting on AI for, and the exposure that creates - **The risk landscape:** how AI risk is evolving for our industry and jurisdictions - **Our governance model:** how it works, what the board sees and decides, what it does not - **Two or three deep dives:** a real system, a real incident or near-miss, a real regulatory change, explained fully - **What good oversight looks like:** the questions directors should be asking management - **Open discussion** ### Worked example **2026 annual board AI briefing:** - Strategy: AI is core to the hiring and fraud products; ~30% of revenue touches an AI-influenced decision. - Landscape: risk-based regulation spreading from the EU to US states; enforcement focus on hiring and lending. - Model: committee under Risk; board sees a quarterly dashboard and any Sev-1; board approves risk appetite annually. - Deep dives: (1) the resume screener end to end; (2) the September support-copilot data near-miss; (3) the EU AI Act high-risk deadline move. - Questions to ask management: "which systems are High-tier and are they all assessed?", "what is our worst plausible AI incident and are we ready for it?", "where are we behind on a regulatory deadline?" ### Acceptance criteria - The briefing is context and education, held separately from operational reporting. - It includes deep dives on real systems, incidents, and regulatory changes. - It gives directors specific questions to ask management. --- ## Director AI literacy self-assessment _A short questionnaire to find gaps and track progress, done privately by each director._ ### Template | Statement | Confident / Somewhat / Not | |---|---| | I can explain, at a high level, how the company's main AI systems make decisions | | | I understand the main ways AI systems fail and which of ours are exposed | | | I know which AI regulations the company is subject to and the key deadlines | | | I understand how AI risk is governed and what the board decides | | | I can read our AI risk dashboard and tell whether the picture is good or bad | | | I know what would trigger an escalation to the board and what we would do | | | I know the questions to ask management about AI risk | | ### Worked example > Aggregated and anonymised across 8 directors, 2026. | Statement | Confident | Somewhat | Not | |---|---|---|---| | Explain how our AI makes decisions | 3 | 4 | 1 | | Ways AI fails and our exposure | 2 | 5 | 1 | | Regulations and key deadlines | 1 | 4 | 3 | | Read the risk dashboard | 4 | 3 | 1 | Gaps to address in the curriculum: regulation, and failure modes. Repeat after the annual briefing. ### Acceptance criteria - Results are aggregated and anonymised, used to shape the curriculum, not to grade individuals. - The assessment is repeated to show progress over time. - Low-confidence areas map to specific curriculum modules. --- ## External educator selection criteria _How to choose an outside educator, and what a good one looks like._ ### Template | Criterion | What good looks like | |---|---| | Board-level experience | has taught directors, not just executives or engineers | | Independence | no conflict (not selling the company AI tools or audit services) | | Tailoring | will build the session around our systems and risks, not a stock deck | | Regulatory currency | current on the regimes we face, with primary-source knowledge | | Practical framing | teaches the questions to ask, not just concepts | | References | directors at comparable companies who can speak to impact | ### Worked example > Shortlist evaluation. | Candidate | Assessment | Outcome | |---|---|---| | A | strong board experience; stock deck only; would not tailor | pass | | B | tailors; current on EU and US rules; references check out; independent | selected for the 2026 briefing | | C | excellent, but also pitching us an assurance engagement | conflict; declined | ### Acceptance criteria - The educator has taught at board level and will tailor to the company's systems. - There is no conflict of interest (not selling related products or services). - References from directors at comparable companies were checked. --- ## Governance controls this kit produces evidence for - **BRD-001**: The curriculum and self-assessment are the director AI literacy and competency assessment. - **BRD-002**: The briefing reinforces the board's role and decision rights under the committee charter. - **HOC-007**: Teaching directors to read the risk dashboard strengthens board risk reporting and escalation. - **BRD-005**: Director literacy is one dimension of the governance maturity assessment.