andrewgcondon-dxrut

Episode 002 · Governance × AI

AI literacy for directors is oversight competence — not coding fluency.

This episode asks what a director actually needs to understand about AI to discharge their duties well. The answer is not “learn to code” and it is not “appoint one AI expert and move on.” It is the ability to ask better questions, understand where AI is being used, challenge risk ownership and govern the board’s own use of these tools.

The master audio is temporarily hosted from the project archive while Spotify distribution is being finalised.

What good oversight looks like

A board does not need to know how every model works internally. It does need enough shared understanding to challenge management on where AI is used, what data it touches, what can go wrong, who owns the decision and where human review or override matters.

• Know the footprint. AI use should be visible across enterprise tools, public tools, the board and the secretariat.
• Insist on ownership. “The model did it” is not an accountability model. Executives still own the decisions and controls around AI use.
• Govern the board’s own behaviour. Board packs, minutes, privileged material and sensitive information create risks when directors or secretariats use public AI tools without clear rules.
• Avoid literacy theatre. A single training session, one AI director or a logo slide is not the same as organisational oversight competence.

Monday morning

Ask for a one-page living map: where AI is being used, on what data, at what level of risk, who owns it, where human oversight sits, and how you would know an AI-assisted output was wrong before relying on it.

AI disclosure: Produced with extensive AI assistance for research, synthesis and production, based on an approved Master Learning Pack drawing on international governance sources including NACD, OECD, NIST AI RMF, ICGN, ISO/IEC 42001 and labelled EU AI Act material. AI use is disclosed. Learning in public — not a claim of original expertise.