Board and executive AI briefing
Boards and executives do not need to build AI. They need enough understanding to ask the right questions, see the real risks through a mission lens, and hold the line on who is accountable. This briefing gives you that footing, so oversight is grounded in judgment rather than headlines.
By the end, you will be able to
- Explain, in plain terms, what AI can and cannot do and who it affects.
- See AI opportunity and risk through the lens of your mission.
- Tell the difference between board oversight and executive operation.
- Insist that affected people can question a decision, not just receive it.
Lesson 1
What leaders need to understand about AI
You do not need the engineering. You need an accurate picture of what these tools do, where they fail, and who stands in the path of a decision.
As a leader, your job is not to master the technology. It is to understand it well enough to govern it. That means a correct, plain picture of what today's AI is and how it behaves.
Modern generative AI produces fluent language and finds patterns in material you give it. It is strong at shaping and summarizing, and it is unreliable about specific facts it was not handed. It can state something false with complete confidence.
- Capabilities: drafting, summarizing, sorting, and surfacing patterns in text you provide.
- Limitations: fabricated facts, uneven performance across groups, and no real sense of its own uncertainty.
- Lifecycle: a tool is chosen, fed data, used in real work, and eventually changed or retired. Each stage is a place to govern.
- Stakeholders: staff who use it, people it makes decisions about, funders who trust you, and the community you serve.
Try it
In one plain sentence, describe what AI is to a fellow board member who has never used it. If your sentence would survive a skeptical question, you have the footing you need.
RememberLeaders govern AI by understanding its capabilities, limits, lifecycle, and the people it affects, not by building it.
Lesson 2
Opportunity through a mission lens
The right question is never whether AI is impressive. It is whether it advances this mission for these people.
AI is only worth using where it serves the mission. Judged that way, the opportunity is concrete rather than abstract. It shows up as more capacity, better learning, and steadier operations, not as novelty.
- Service capacity: reaching more people, or serving them better, with the same staff.
- Learning: understanding what is working in programs and what is not.
- Resilience: handling surges, gaps, and disruptions without dropping care.
- Fundraising: freeing staff time for real relationships rather than routine drafting.
- Operations: reducing friction in the background work that keeps the doors open.
Note
Hold each opportunity to a simple test: does this help us serve our mission better, for the people we exist to serve? If the answer is only that it is new or fast, it is not yet a reason.
Try it
Name one place in your organization where AI could genuinely expand mission capacity. Then name what you would need to see to trust it there.
RememberOpportunity is real only when it advances the mission for the people you serve, not when it is merely impressive.
Lesson 3
Material nonprofit risks
The risks that matter for a nonprofit are not the science-fiction ones. They are mission drift, lost trust, and harm to the people you serve.
Every organization faces a specific set of AI risks that a board can actually watch for. These are the ones with weight, because they touch the mission, the people, and the trust that holds the organization together.
- Mission drift: chasing what the tool makes easy instead of what the mission needs.
- Public trust: a visible error or a misuse that erodes the confidence of donors and community.
- Unequal access: a tool that works well for some people and quietly worse for others.
- Sensitive data: exposure of information about beneficiaries, donors, or staff.
- Workforce impact: change that lands on staff without their voice in it.
- Vendor dependence: relying on a tool you do not control and cannot easily leave.
Important
Consider a hypothetical: a service tool that quietly works better for English speakers than for the families you serve in other languages. No one chose that harm, and no one is watching for it. That is exactly the risk a board exists to surface.
Try it
Of these six risks, which is most alive in your organization right now? Naming the live one is where oversight starts.
RememberWatch the risks with mission weight: drift, trust, unequal access, sensitive data, workforce, and vendor dependence.
Lesson 4
Board oversight versus executive operation
The board does not run the AI. It decides what must be decided, monitored, delegated, or declined. Keeping that line clear is the work.
Good governance depends on a clean line between the board's role and the executive's. The board sets direction, asks hard questions, and holds accountability. The executive team operates within that direction and reports back.
- Decide: the board owns matters of principle, risk appetite, and mission alignment.
- Monitor: the board watches agreed measures and incidents through regular reporting.
- Delegate: the board hands operational choices to the executive, with clear boundaries.
- Decline: the board can say some uses are not for this organization, full stop.
Try it
Take one AI use your organization is considering. Sort it into decide, monitor, delegate, or decline. Notice how much clarity that one sort brings.
RememberThe board decides, monitors, delegates, or declines. Accountability stays inside the organization, with a named human.
Lesson 5
Stakeholder legitimacy
A decision the affected people cannot question is not legitimate, however efficient it is. Voice and appeal are part of governance, not extras.
For a mission-driven organization, legitimacy comes from the people you serve. That means their voice belongs in how AI is used, and it means anyone affected by a decision can find out, understand, and challenge it.
This is not softness. It is the difference between a system that serves people and one that processes them. When beneficiaries can participate and appeal, errors get caught and trust is kept.
- Beneficiary voice: the people served have a real say in decisions that affect them.
- Community participation: affected groups are consulted before a use is rolled out, not after.
- Correction: a wrong decision can be found and fixed.
- Appeal: a person can question a decision and reach a human who can change it.
Try it
Think of one AI-supported decision that would touch the people you serve. How would one of them question it, and who would answer? If you cannot say, that is the gap to close.
RememberLegitimacy requires beneficiary voice and a real path to correction and appeal, not just an efficient decision.
What you leave with
Draft AI principles, baseline risk posture, and board question set
A short leadership package in your own words. It holds a first draft of your AI principles and an honest read of where risk sits today. It also lists the standing questions a board should ask about any AI use. It gives your board a shared starting point instead of a blank page.
This is just for you. It saves on this device only, and nothing is scored.