Here’s how to create an AI policy for a nonprofit that your team can actually use. Start with three decisions staff face in real work: which tools they can use, what data they can share, and who checks the results.
We spoke with Dr. Stephanie Rose-Belcher, chief operating officer at JMT Consulting, about making those decisions for fundraising, finance, and everyday administration. The conversation opens with a donor CRM intrusion and turns to questions leaders can bring to their next software renewal. What AI features are already built into the product? Who controls the settings? How does the vendor use the information your staff puts into the system?
Rose-Belcher walks through three kinds of tools to account for in a nonprofit AI acceptable use policy: personal AI accounts, organization-managed AI licenses, and AI features embedded in software your team already uses. That third category deserves a look anywhere staff handle donor records, accounting information, or payroll. As she puts it, “Technology is changing too fast for us to write a policy and stick it in a drawer.”
Then comes the bank reconciliation. AI can flag discrepancies. The staff accountant investigates and documents the work, and the controller reviews and approves it. Rose-Belcher uses that example and a familiar one, automated meeting notes, to show how the four RACI roles work: responsible, accountable, consulted, and informed. Her rule for meeting notes is refreshingly direct. “I am accountable for bad notes going out. That’s on me.”
Scan the on-screen QR code for JMT Consulting’s nonprofit AI policy checklist and use it to decide what your team will allow, what information staff can enter, and who reviews the result.
Key Takeaways:
– Write a short purpose statement that names the work AI should help your team do.
– Maintain an approved tool list that includes AI features inside existing software.
– Review vendor terms for data storage, retrieval, and use in model training.
– Assign a person to check AI-generated work and specify who gives final approval.
– Start with one routine workflow and update the policy as tools and uses change.