Digital and AI Literacy as a Governance Competency – Part 6

Uncategorized

Part 6 of 8 — Director Education: An Association Imperative

A staff member walks into the boardroom with a slide deck and an AI vendor proposal. Half the board nods along without asking a single hard question, because the topic feels technical and they don’t want to look foolish. The other half pushes back reflexively; not because they’ve identified a real risk, but because the whole idea makes them uneasy. Neither group is governing. Both are avoiding the work.

That avoidance is now expensive. Technology and AI decisions touch member data, financial exposure, reputational risk, and the association’s basic capacity to operate—which makes them governance questions, not IT questions. I am very cautious about using any AI functions. However, that is not the subject of today’s post; it’s just a note that you should always ask the hard questions before using any AI functionality. The duty of care every director already owes doesn’t pause when the topic turns to software or artificial intelligence. Still, that duty was built to prevent the most obvious harms, not to produce sound technology judgment—everything above that floor is a matter of disciplined curiosity, not a legal minimum. A board that outsources its judgment on these decisions to whichever staff member sounds most confident has quietly stopped governing that part of the organization.

Why boards default to rubber-stamping or obstruction

Most directors arrive at the table as experts in their profession, not in data architecture or machine learning. Faced with a proposal that uses unfamiliar language, a board has two easy exits: defer entirely to staff, or resist on instinct. Both responses feel safer than admitting uncertainty out loud. Both also abandon the board’s actual job.

The deferral pattern is more common, and more dangerous. It looks like diligence: a vendor demo, a few polite questions, a vote, but it is really the board asking to be told what to think. The obstruction pattern is rarer but just as costly: a director blocks or slow-walks a legitimate initiative because the unfamiliar feels risky, and the organization loses ground to peers who moved. Neither pattern requires directors to become technologists. Both require directors to learn a different kind of question.

Governance questions come before feature questions.

Here is the reusable filter: governance questions come before feature questions. Before any director asks what a tool does, the board should ask what it’s for, what it risks, whether the organization can actually run it, and who answers for it when something goes wrong.

Four questions, asked in this order, on every technology or AI proposal that reaches the board table:

Value — What member or organizational problem does this solve, and how will we know if it worked? Not “is this innovative,” but “what changes for members or operations if we approve this.”
Risk — What data does this touch, who else can access it, and what happens if it fails or is misused? This includes reputational risk, not just technical risk.
Capability — Do we have the staff, budget, and process maturity to actually operate this responsibly, or are we buying a tool we don’t have the discipline to run?
Accountability — Who owns this once it’s approved: which staff role, which vendor, and which board committee monitors it going forward?

Only after those four are answered does “what does the software do” become a fair question. A board that reverses that order ends up evaluating a demo instead of a decision.

The literacy directors actually need

None of this requires directors to code or to understand model architecture. It requires a working, conceptual grasp of four areas:

Data stewardship and privacy basics. What member data does the association collect, where does it live, and what obligations attach to it? Directors don’t need to design a data policy; they need to recognize when one is missing.
Cyber-risk exposure. Every new system is a new attack surface. Directors should know whether the organization has basic protections in place and whether a breach response plan exists, not the technical mechanics of a firewall.
AI bias, transparency, and IP considerations. AI tools trained on unrepresentative data can produce skewed outputs, and content generated by or fed into AI systems raises intellectual property questions the association may not have addressed. Directors don’t need to audit a model: they need to ask whether anyone has.
Realistic readiness versus technology theater. Some initiatives look impressive in a board presentation and have no real operational discipline behind them; no trained staff, no maintenance plan, no adoption strategy. Directors need the judgment to tell the difference between a genuine capability and a demo dressed up as one.

This form of Director Literacy is the essence of governing for what is next rather than managing what is in front of you: a compass for judging technology decisions, not a manual for building them.

Why this is now a structural question, not just a behavioral one

The stakes go beyond any single vendor vote. Associations are increasingly restructuring their governance infrastructure around digital-first participation: how members engage, how input is collected, and how decisions are made increasingly run through online systems as the default rather than the exception. Governance bodies built for an era of in-person representation are being redesigned for one where digital access shapes who participates and how.

The American Counseling Association offers a carefully bounded illustration of this structural shift. According to ACA’s own governance materials, the association is moving from a 32-member representational Governing Council to a smaller, elected Board of Directors under bylaws revised in 2025, with a new Advisory Council taking on a broader representative and advisory role (ACA’s new governance model). ACA frames this as a deliberate modernization of how the association is governed, not a routine bylaws update (ACA bylaws, June 2025). The specifics of ACA’s transition are its own, and this post makes no claim about ACA’s internal motivations beyond what its published materials state. The broader pattern worth naming is this: as associations redesign who governs and how, the technology that carries member input, oversight, and decision-making stops being a back-office concern and becomes part of the governance structure itself. A board that hasn’t built fluency in that terrain is not just behind on a tool; it may be behind on understanding its own operating structure.

An AI oversight question set for the next vendor proposal

The next time a staff proposal includes an AI tool or vendor, a board doesn’t need a technical audit. It needs sharper questions, asked before the vote:

What data does this tool use, and does it include member or personally identifiable information?
Has the vendor been asked directly about bias testing, and what did they say, not what does their marketing claim?
Where in the process does a human review the AI’s output before it reaches a member or the public?
What happens to our data if we terminate this vendor relationship: can we get it back, and can we make them delete it?
Who inside the organization is accountable if this tool produces a harmful or embarrassing error?
Is this reversible? If it fails in six months, what does unwinding it cost us, financially and reputationally?
What did we decide not to ask the vendor, and why?

A board that can answer those seven questions has governed the decision. A board that skips straight to “does it save staff time” has evaluated a feature, not a risk.

Where this leaves us

No director needs to become a technologist to do this work well. What every director needs is the discipline to ask governance questions before feature questions, every time, without exception, regardless of how confident the presenter sounds or how urgent the timeline feels. That discipline is learnable, teachable, and, like every other capacity in this series, decays without practice.

Pull up the minutes from your board’s last technology or AI-related vote. Read what was actually asked before the vote was called. Was it governance, or was it deference dressed up as diligence?

If you want a structured way to find out where your board’s technology and AI oversight actually stands, a Future-Ready Board Assessment will show you the gap in plain terms, or bring this filter to your next board meeting yourself: before the next AI or technology item hits your agenda, insist the discussion open with value, risk, capability, and accountability, in that order, before anyone shows a single feature. If you’d rather talk it through first, start with a Strategic Capacity Briefing.

Let me know what you think. Next week: Part 7- Adaptive Leadership in the Boardroom: Getting on the Balcony, Regulating Disequilibrium, Building Coalitions