Mission-driven AI Case Studies: New Series!
Oops! Another series.
I’ve noticed that a lot of the implementation examples we all talk about (both as aspirational and cautionary tales) are in for-profit contexts, but there’s lots of instructive work going on in mission-driven organizations, too!
Profit-driven organizations focus primarily on efficiency when implementing AI; mission-driven organizations need to care about efficiency, but also ensuring it advances rather than undermines their core mission and values.
In this series, we will learn:
What does AI implementation look like across different organizations? We’ll see first hand how organizations with different sizes, missions, and sectors implement AI.
What capabilities of AI do these implementations showcase? What tasks are they automating, and how are they mitigating (or could they mitigate) risks in their implementation? This will extend beyond LLMs and give us a view of how many types of AI can supercharge your mission.
What ethical and effectiveness concerns should you be considering for your similar implementations? Through real-world examples, practical frameworks, and actionable advice, this series will help your organization think through how you might leverage AI.
What results have they seen so far? Efficiency? New services? Scaled impact? Total disaster? We’ll see how AI has impacted the missions of organizations who have implemented it and what we can learn. Learning the easy way is my favorite of all the ways!
As we explore these implementations together, we'll build a understanding of what responsible, mission-aligned AI can look like in our sector. I look forward to hearing your questions, experiences, and challenges as we navigate this journey.
Our first post, out tomorrow, covers how GiveDirectly is using AI to help identify high-need areas after natural disasters. Let me know if there’s an implementation story you’d like me to dig into, or if you’d like to share what you are doing!
If you aren’t using LLMs right now but you’d like to start, download your free Mission-First AI Starter Kit and check out the Safe Start Prompt Pack!
—
LLM disclosure:
I asked Claude 3.7 Sonnet:
”I am writing a blog series about how mission-driven organizations are using AI. I think these examples will help readers see how AI can be used well and how they can avoid pitfalls. I am writing the introduction post for this series right now, and I've attached what I have so far. Can you give me some more ideas about how I can make sure this content will help mission-driven organizations and the people who work there? I want to make sure this series is useful. I am interested in 1) instructions for me (as the author of the series) to make sure that the series is useful and 2) an idea of how to communicate the value of the series in the intro post.”
Overall, I thought the result I got with this query was not that helpful. The model assumed that I would have a lot more insider details about implementations than I expect to get, so a lot of the text it spit out wasn’t useful.
I also asked ChatGPT o3 to make an image for this by pasting in just the text (not the LLM disclosure, to save some resources and avoid swaying the context).