How we’re thinking about generative AI

Originally published October 2023

Whenever I encounter a Zoom call malfunction or some other IT issue, I like to make the same joke: “We’re data people, not tech people!” And while that’s often true — after all, our sector isn’t known for being earlier adopters — some emerging trends can’t be ignored. Generative AI is one of those trends.

As such, the Nexus team has spent more time than we expected thinking about generative AI and its implications in our sector. Honestly, my thinking is usually to figure out what AI can do and run the other way, toward the hard-to-do human-centered data work that machines aren’t good at (yet). But there is a lot of opportunity here, and we should think strategically about it.

So, here are five high-level thoughts we want to share with a few thousand of our closest friends:

1. Values. Like anything else your organization does, your core values should drive any and all behaviors: hiring, policies and procedures, program design, fundraising, everything. AI isn’t any different. When and if you’re thinking about incorporating generative AI into your organization, you should carefully consider how it aligns and strengthens your organization and those you serve. Sometimes the best use of AI is not to use it. Internal guidelines can keep you on track.
2. Idea generation. The easiest thing to do with AI, and what you’ve probably seen the most of, is generating lots of new ideas. Using something like Chat GPT, you can generate a lot of ideas and content. To get the best results, Google “prompt engineering tips.”
3. Replace rote tasks. From note-taking to data entry/transcription to a “competent but naive assistant,” look for ways AI can automate your most tedious processes. We’ve even used it to translate blog posts. This can save your human staff more time to do work that humans are best at.
4. Copilot. The AI opportunity in nonprofits and other social impact organizations is likely more as a copilot than providing services itself. I would be extremely hesitant to let AI directly interact with clients or program participants, like case management, customer support, etc. But that doesn’t mean it can’t help your staff find information and make decisions as they’re serving people–even in real-time.
5. More complex data analysis. AI is allowing us to combine and analyze larger and more complex data sets. For instance, if you were to create a chatbot for your audience to interact with, you’d want to train it on a closed data set rather than the open internet. Generative AI might be the chatbot’s interface, but it could also help you create the underlying dataset by scraping information from myriad spreadsheets, PDFs, even videos. Another example: You know how painful it is to analyze open-ended survey questions at a large scale? Well, improved natural language processing can help us ask those types of questions at scale and provide better insights than simplistic word clouds.

👉 The bottom line: Be curious. Be creative. Be skeptical. Be values-driven. Let AI save you time, but don’t let it make decisions for you. Generative AI isn’t magic–it’s a tool that you can leverage for your organization.

Where am I learning about Gen AI? Read The Smart Nonprofit. Follow Beth Kanter on LinkedIn. Follow the Atlanta Interdisciplinary AI Network (AIAI). Check out online and in-person learning opportunities. Listen to literally any business or tech podcast (the jargon will eventually start to make sense).