Report on the state of AI in non-profits
I’ve read an incredible and massive report from Make Sense Of It on the state of AI in non-profits, and summarized 12 key takeaways that seemed the most important for me.
Link to the report: https://www.aiplaybookforcharities.com/2026
1. The efficiency gap. 40% of leaders save 8 or more hours a week using AI, while two-thirds of staff members save less than 2 hours
2. The sustainability of AI is becoming a pressing issue due to the massive resources consumption of data centers. If you care about your carbon footprint, using AI for casual chatting or trivial tasks is something to reconsider
3. Invest in expertise, not just tech knowledge. AI can execute a job perfectly, but it cannot define the problem itself. Success lies in a deep understanding of your domain and your team's ability to "explain" that context to the AI
4. The trust factor. The closer AI gets to the external world, the stricter the guardrails must be. Trust is an NGO’s primary currency. Systems interacting with your beneficiaries require far more rigorous testing than internal tools
5. Transparency matters. Always inform people when they are interacting with an AI representing your organization. Avoid misleading your audience, even unintentionally
6. Garbage in, garbage out. Bad, unstructured data is a major risk. AI won't clean up your messy data; it will just confidently scale those errors across its outputs
7. Work isn't shrinking; it's shifting. AI handles the mechanical tasks. Spending human hours on routine work (like scheduling or basic translation) is now wasteful. Reinvest that freed-up time into tasks that require human judgment and empathy — areas where AI shouldn't belong
8. Change management over tech. Implementing AI is about change management and restructuring workflows, not just deploying software. If you inject AI into broken processes designed for humans, you'll just get faster, more broken processes
9. Prevent "Shadow AI". Provide your team with official, secure, and paid AI tools. If you don't offer convenient, approved options, employees will use free versions in secret, risking leaks of sensitive data
10. Treat AI like a new intern. It needs clear onboarding, constant supervision, strict boundaries, and rigorous quality checks. But, unlike a human intern, AI doesn't genuinely learn from its context; it will just keep making mistakes with absolute confidence
11. "Should we?" vs. "Can we?". Just because AI can automate beneficiary outreach or streamline volunteer management doesn't mean it should. In the social sector, the core value often lies entirely in human connection
12. Measure what matters, not what's easy to track. AI makes it easy to crunch massive data sets, but that’s no excuse to invent vanity metrics just because you can. Experienced staff must define the indicators that truly measure social impact, using AI strictly as a tool to gather them.
