Buolamwini.When we were approached by leadership of AFP Colorado to host a bookclub around ethical fundraising and the hottest topics of the moment, AI immediately rose to the top of the list. 

For those of us who have been in the sector for a minute, we can likely remember going from a physical Rolodex to a CRM (likely by way of a spreadsheet first). My first nonprofit job had a Microsoft Access database that was glitchy and constantly crashing. 

For a sector that’s constantly behind in technology adoption and access, getting AI right is not only important, but it’s likely existential. At the same time, as nonprofits, we know to approach with care, caution, and leading with our values. That’s what led us to read “Unmasking AI: My Mission to Protect What Is Human in a World of Machines” by Joy Buolamwini.

Over the course of 4 sessions with a cohort of fundraisers, we explored the basic mechanics of how fundraisers are using AI, the limitations of AI systems, and the ethical implications. Some of the biggest takeaways from our discussions: 

1. Fundraising will always require the human touch. Although we can (and likely should) use AI to improve efficiency, we must ensure everything is reviewed by a person. AI can help us focus more on developing deep and meaningful relationships with donors and the community, while allowing robots to take care of the back office work. 
   
   
2. We must be cautious of what information we feed into AI systems. When they’re “free,” it’s because the data we provide is what is paying them. Anything that is identifying, confidential, or otherwise shouldn’t be released to the world should never be fed into an AI System. 
   
   
3. The ethical imperative is to ensure that we are always aware of the biases inherent in AI systems. They have been trained on biased data that will bias their outputs. Here’s another place where humans must intervene to provide oversight and address these issues.