Skip to content
EP 682March 18, 2024

Artificial Intelligence For Nonprofits, Redux

Tony MartignettiJustin SpelhaugAmy Sample WardTristan Penn
Nonprofit Radio
Episode 682
0:00 / 0:00

About this episode

A second savvy panel takes on the impact, leadership demands, promises, responsibilities, and future of AI across the nonprofit community. We convened a panel in June last year. But this is an enormous shift in nonprofit workplaces that deserves another look. This panel is Justin Spelhaug, from Technology for Social Impact at Microsoft, and Amy Sample Ward and Tristan Penn from NTEN.

Here’s the transcript

Tony MartignettiHello and welcome to Tony Martignetti Nonprofit Radio, big nonprofit ideas for the other 95%. I'm your aptly named host and the podfather of your favorite hebdomadal podcast. Oh, I'm glad you're with us. I'd be forced to endure the pain of chronic inflammatory demyelinating polyradiculoneuropathy if you attacked me with the idea that you missed this week's show. That one is so good, it deserves 2 weeks. And plus, I spent a week practicing it, so it lives on for one more week. Here's our associate producer, Kate. To introduce this week's show.

Kate MartignettiHey, Tony, I'm on it. It's Artificial Intelligence for Nonprofits Redux. A second savvy panel takes on the impact, leadership demands, promises, responsibilities, and future of AI across the nonprofit community. We convened a panel in June last year, but this is an enormous shift in nonprofit workplaces that deserves another look. This panel is Justin Spelhaug from Technology for Social Impact at Microsoft, and Amy Sample Ward and Tristan Penn from NTEN, on Tony's Take Two.

Tony MartignettiThank you.

Kate MartignettiWe're sponsored by DonorBox. Outdated donation forms blocking your supporters' generosity? DonorBox, fast, flexible, and friendly fundraising forms for your nonprofit. donorbox.org. And by Virtuous. Virtuous gives you the nonprofit CRM, fundraising, volunteer, and marketing tools you need to create more responsive donor experiences and grow giving. virtuous.org. Here is Artificial Intelligence for Nonprofits Redux.

Tony MartignettiWe're talking this week about artificial intelligence again. It's an important topic. Uh, we did this with a panel in June last year. Today, a different distinguished panel shares their thoughts on this transformative technology. It's timely. It's got a lot of promise and a lot of risks. It's moving fast. Those are the reasons why Nonprofit Radio is devoting multiple episodes to it. What are the promises and the responsibilities? What's the role of nonprofit leadership? How about government? What are the equity concerns, the biases? What about access to this intelligence? What are the preconditions for successful integration at your nonprofit? What's the future of artificial intelligence? Here to share their thinking are Justin Spelhaug, recently promoted Justin Spelhaug. He is Corporate Vice President and Global Head of Technology for Social Impact at Microsoft. You'll find Justin on LinkedIn. Justin, welcome to Nonprofit Radio. Congratulations on your promotion from Vice President to Corporate Vice President at the enormous company Microsoft.

Justin SpelhaugIt's great to be here with the Podfather. It's a new name, so I'm proud to be here and look forward to the conversation.

Tony MartignettiAll right. Well, I'm glad it's the first time you've heard the Podfather. There can be only one, really. There ought to be only one. So I'm glad it's the first time. And I see global head. I'm sorry you're a little bit limited. You're not working in the stratosphere, the ionosphere, the troposphere. You're strictly limited to the globe. I'm sorry. Yeah. We all have our constraints.

Justin SpelhaugWe are working on Mars and the moon soon, but we gotta get a broader population of nonprofits there.

Tony MartignettiAll right, so we're limited to the globe. I'm sorry for you. Amy Sample Ward, we know them. They're Nonprofit Radio's technology contributor and the CEO of NTEN. They're at amysampleward.org and @amyrsward. Amy, it's great to see you. Welcome back, of course.

Amy Sample WardThanks. I know there have been a number of different conversations about AI that you've had on Nonprofit Radio. Um, I've listened to them, haven't been in all of them. They've been great. And, you know, we talked a little bit about AI, Gene and I, you know, when we started off with some of what's going to be big topics in the sector for 2024. So excited to be in a conversation kind of dedicated to that.

Tony MartignettiI'm glad you are. And Tristan Penn, welcoming back Tristan. He is Equity and Accountability Director at NTEN. As a Black and Navajo professional, he's served on previous organizations' equity teams and been a facilitator for DEI rooted in racial equity. Tristan is on LinkedIn. Tristan, welcome back.

Tristan PennAwesome. So happy to be here. Thank you for having me. Excited to have this conversation with Amy, who I work very closely with. And it's really good to see you too. And also excited to have this conversation with Justin to see, you know, what we can unearth.

Tony MartignettiYes, we're representing the big tech perspective. Amy, since you are our tech contributor, we're going to start off just big picture. What are your thinking? What is your thinking? What are your concerns? Big picture stuff.

Amy Sample WardYeah. Well, I'm glad that we've scheduled 5 hours for this interview. I will be taking the first 4. Thank you so much. I have many thoughts, many concerns, many— I think there's just a lot to get into. I think some top-level, you know, bites to put at the beginning here are there's a lot of hype. And as with anything that falls into the hype machine, I think nonprofits do not need to fall, you know, victim to like, oh my gosh, I read this one article, so I have to do the thing, right? There's, there's time. Mm-hmm. AI is not done. The world is now, now not already over and everything's predetermined, right? So You've seen the article that was like, AI will end humanity. Okay, okay, here we are. Let's calm down and talk about things. So I, I know I've talked to nonprofits whose boards are like, I read that article and AI is good, you know, it's ending all of us. Like, we can take our time. That's one piece. I also think it's important for organizations to think about where they are already working, what communities they already work with, what data they already have. Like, this isn't start a new project when we're talking about AI. And so I think we'll get into that more in our, in our conversations here. And of course, that AI isn't new. Well, I mean, artificial intelligence as a phrase is the broadest umbrella term we could use for these types of technologies. And so To, to have these sentences that say like, AI is new and it's here and it's going so fast. Like, what does that— that's like encompassing so many different components of technology. And so what do, what do we really mean when we're talking about AI? Are you talking about a model that you set up inside of your organization, you know, to help identify program participants that need extra support? That could— that can be AI, but that's very different than saying Oh yeah, we're just using ChatGPT to help, you know, start some of our drafts. Okay, those are so— they're wildly different things. And so to talk about them in the same breath as it's all AI, it sets folks up to already have kind of a disconnected conversation even from the start.

Tony MartignettiAll right, thank you. And hold our feet to the fire, especially me, because the 3 of you think about this all the time. And I don't. So, you know, if I, if I lose that context that you just revealed, shared with us, please call me out. All right, Justin, big picture, please. Yeah, what do you think?

Justin SpelhaugAnd I'll build on Amy, you know, the hype cycle of it's gonna save us, it's gonna destroy us, and now just kind of how do we make use of it? We've been going through this, this process as a, as a community. I, I think one of the things when I zoom out, I, I just see Some tectonic shifts that are impacting the sector from some big demographic shifts in European countries, in the United States, where our workforce is getting older, that's putting tons of pressure on aged care and frontline community workers. Some big shifts in continents like Africa where education, skilling, and jobs are all critical. And the nonprofits facing off on these issues aren't getting any additional funding. GDP has stabilized in many countries, but we've hit a new set point for inflation that's impacting pocketbooks. It's impacting people's ability to raise money. And so really, you know, the question that we have to ask is, how do we use AI in, in missions to help organizations raise more money, help them deliver more effective program, help them rise to these challenges that are continuing to create pressure in the sector? And how do we do all of that in a way that's responsible, in a way that's safe, in a way that's inclusive? And that's actually a pretty complex topic that I hope we spend some time on.

Tony MartignettiIndeed. And thank you for the global perspective. Tristan, big picture thoughts, please.

Tristan PennI have lots of thoughts similar to Amy. And I think where I start off with is kind of like in a very— I worked for 20 years and I still am working in nonprofit. And I see how over those years nonprofits and, you know, small organizations have seen something that's bright and glittery and been like so amazed by it been like, yes, we want it, we're going to take it in, and we have no process for building it into our operations. We have no forethought for it. We have no contingency to live by when we're folding this in, this ideal state. We've already jumped to multiple steps to us envisioning how we're going to operate with this bright, shiny tool that we have. And that's never been the case in my years that I've I've been in nonprofit and it's, if anything, it's always been folded in in a way that doesn't have a lot of forethought too. So I think the things that come to mind for me that make me curious and also a little bit reticent about just the blanket, the umbrella term AI is folding it in where it makes sense and not where you want to add a little, you know, icing on your cake where it does— where it needs none. And so that's where I intersect with it. There's another piece of it where I am a little critical of it and concerned about it because I think that this can, you know, we, to Amy's point, we think about AI and a lot of people go in different directions. I think the baseline for a lot of people is they go to like AI-generated pictures or ChatGPT to do those things. And it's much more than that. But I do think about a time anecdotally where I was at a conference and I was passing by a booth and there was like a very lovely, you know, picture of an older couple. And I was like, oh, that reminds me of my grandparents. It was an older Black couple. And I was like, oh, that's so cute. It reminds me a lot of my grandparents. It's like very, you know, and then I went in closer and this is a booth that's, you know, managed by a bunch of white folks. And then they were like, oh, did you know that this is an AI-generated picture? And that didn't feel good To me as a Black person, that didn't feel good. It felt incredibly like I had been misled in a really scary way. I feel like I have a really good detector of like what's real, what's not. My BS detector is like always up and on. And that scared me because I was duped hard. And that scares me in a way less about nonprofits, but just the overall globalization and usage of it and implementation that it could go into the hands of people and create false narratives about marginalized groups. Yeah. Just based on what they what product they want to sell, and that is scary. And that that's something that I think has just stuck with me for for a while.

Tony MartignettiThank you for raising the the risks and and potential you know misuse abuse. We we need to go to artificial intelligence to create a a picture of an elderly black couple. Right? That was it was necessary to do. And also, thank you for the valuable parallel. You know, you you make me think of. Social media adoption when Facebook was new. You know, we assigned it to an intern and we put it like the cherry on top where we didn't need a cherry, but the intern had used it in college. So, you know, she may as well do it for us full time. Very valuable, interesting parallel. Amy, start us off with just a common definition. You know, Artificial intelligence, generative. I mean, generative artificial intelligence. That's what we're largely going to be talking about, if not exclusively. I think so.

Amy Sample WardYeah.

Tony MartignettiWhat is generative?

Amy Sample WardYeah, there's a lot of terms that I think we're—

Tony MartignettiWell, we'll start with taking one at a time, right?

Amy Sample WardSure. No, I was just going to say, I think we already are exposed when we're thinking about technology in our nonprofit organizations to lots of different terms, lots of different companies putting things out there with the, uh, not necessarily cloaked, you know, it's not, it's not a hidden desire to reinforce that they're specialists, they know what they're doing. And like us lowly nonprofits don't know, we couldn't understand those fancy terms, right? And so I always, I mean, I teach a course in AI and always remind folks like, you absolutely can know what these words mean. You know, and I appreciate that there are so many places, even actually, like, I'm never somebody that promotes these things, so folks know this, but like Microsoft has actually offered, you know, community learning spaces to say these are what these words mean. So artificial intelligence is, like I said, the biggest umbrella term for all different types, generative AI, Machine learning, all of these components that people might talk about as if they're one different thing, they're all like within that same AI umbrella. And I just want to say 2 words because they'll probably come up in our conversation. I know you want to go one word at a time, but the words I hear from folks the most where they're not— they feel like they should know what this word means and they don't, and they feel like silly that they don't understand— are algorithm and model. Yeah. Those words are used all the time in talking about generative AI, which means the tool is, is set up to generate something back for you. Tristan used an image, you know, visual image example, but that could be text, that could be video, that could be audio. You know, it's asking the tool to generate something for you. But an algorithm, we've heard this word like, you know, oh, Facebook's algorithm is like choosing what I see, right? The algorithm means the set of rules. So in Facebook's newsfeed, that set of rules says if something already has a bunch of likes, prioritize it, right? If it has, you know, 2 friends that you're connected to already commenting, prioritize it. So it's whatever that set of rules is that says this is how to generate an older Black couple image. Whatever those rules were, that's what algorithm means. And model essentially means, like, you can think of the same— the word is used in the same way As when you say model about cars, like it is the whole set put together, right? It's got the data, it has the algorithm, the rules that say how to do it. It has the input, whatever you're gonna ask it to do. That kind of, when people say what's the model, they're really saying, okay, what's the package of how this tool is working?

Tony MartignettiThank you for all that.

Kate MartignettiIt's time for a break. Open up new cashless in-person donation opportunities with DonorBox Live Kiosk, the smart way to accept cashless donations anywhere, anytime. Picture this: a cash-free, on-site giving solution that effortlessly collects donations from credit cards, debit cards, and digital wallets. No team member required. Plus, Your donation data is automatically synced with your DonorBox account. No manual data entry or errors. Make giving a breeze and focus on what matters— your cause. Try DonorBox Live Kiosk and revolutionize the way you collect donations in 2024. Visit donorbox.org to learn more. Now back to Artificial Intelligence for Nonprofits Redux.

Tony MartignettiJustin, I see you taking lots of notes. What's, uh, what's going on? What's going on in your head? What—

Justin SpelhaugNo, I think, um, what— just as Amy highlighted, one of the things that's important to highlight is, um, we've been using AI for a really, really long time, and there are really important use cases that have nothing to do with generative AI. Things like machine learning, right? That allows us to do things like predict donation propensity. Things like machine language that allows us to translate from one language to another, things like machine vision that allows us to identify and classify objects. All of those are important tools as we look to solve different problems in the sector. Generative AI is, as Amy was highlighting, is a new class of artificial intelligence that allows— that's capable of creating Effectively novel content because it's reasoning across, you know, all of the information in the internet and using, as Amy was highlighting, algorithms to identify patterns that allows it to, you know, produce answers in a really, in many times, intelligent ways. However, as Tristan was highlighting, you know, ensuring that these models are inclusive, are representative, are safe, are understood. are all things that we're continuing to work to put frameworks around and tools around so that they produce positive impact, not negative impact.

Tony MartignettiAnd Justin, how can we ensure that that actually happens? You know, there's a lot of talk about biases. You know, the large language models are trained on predominantly white language sources. So you're gonna, there's, so there's bias. So that, you know, there are equity issues. But what is the big tech doing to actually keep these— keep equity centered in and keep lack of biases centered as these models are adopted using the algorithms that Amy just defined for us?

Justin SpelhaugYeah, it's a really multifaceted answer. I'll only hit 2 points and we can go much deeper if we want. What are you doing about that? We released just, in fact, in the last week, this, the Microsoft AI Access Principles, trying to get at this very problem, which has 11 core components. I'll speak to 2 to give you a flavor of the kinds of things that we need to do as we think about the AI economy globally to ensure it's fair, representative, and safe. One of the principles is making sure that AI models and development tools are broadly available to software developers everywhere in the world, everywhere in the world, in every culture in the world, training on the language and on the history and on the societies all around the world to create much, much more representation. As you probably know, many of the models have been developed in North America and therefore reflect some of those cultural biases. So federating these tools out is critical in the, in the AI economy. Secondly, you know, companies and organizations that produce AI need to have rules for how they check and balance the AI to ensure that it's responsible, it's fair, it's safe, it respects privacy, respects security, it's inclusive, it's transparent. And we call those rules at Microsoft our Responsible AI Framework. And it's not just a set of principles, it's actually an engineering standard. And when applying that engineering standard, we were looking at fairness in speech-to-text. So taking speech and transforming it into text. And we found— it was a couple of years ago we produced this article— that our speech-to-text algorithms were not as accurate for Black and African-American communities in the United States as they were for Caucasian communities. And that was largely a function of the training data that was used. And so we had to take a step back using our framework that caught this issue to say, how do we work with the communities more effectively? How do we bring sociolinguists in to help us understand how to capture all of the rich diversity of language to make sure that our speech-to-text capability is representative of every citizen that we're rolling this out to. And that's an example, and we did that. And today it performs much better and there's more work to do, but it's those kinds of frameworks and guardrails that are really important in helping people design this stuff in a way that benefits everyone.

Tony MartignettiTristan, what's your reaction? You're thinking about equity all the time. What do you say to that?

Tristan PennYeah, what's my reaction? What isn't my reaction? And I would say I love that. And I love what Justin was saying about how, you know, making it a federated model as opposed to it— I mean, yeah, everything— I won't say everything, but a good amount of things are being generated, created, curated in North America and baked into those models. And algorithms are biases that skew towards white men. And that's not okay. I think that excludes me in particular, but also, I think having a plan for that as opposed to being reactionary to being like, well, gosh, we didn't know what was going on, and being a little more less reactionary and more forward thinking in that way. Proactive. Yeah, proactive is always a good place to start. I think a few other things that do come to mind too in terms of making sure that communities of color, marginalized communities are not constantly shouldering, even outside of AI, but constantly shouldering the mess-ups of like the brand new tool that came out on the market. And that seems to always be the case. And there's always like a headline months later where it's like, so-and-so, we found out this tool wasn't geared towards, or facial recognition wasn't geared towards like, you know, Black folks. And it was like, Historically wrong. And so I think about those things, but I also think about it through a nonprofit lens because we're on a nonprofit call. And I bring up another anecdotal story of having— being on a call and having an AI note-taker bot hop into the Zoom call too. I think we've, within like the last half year, we've been on calls where it's like, oh, I don't know if that's an actual person or a thing. Or like, you know, it's very ambiguously named sometimes where it's like—

Tony MartignettiOtter. One of them is Otter, right?

Tristan PennYeah, right. Right. And

Tony MartignettiI— this otter is all of a sudden it's in our meeting, this otter.

Tristan PennRight. Yeah. And I think, you know, there's a lot of benefit. There's a lot of benefit in, in having, you know, note-taking tools and also captioning tools that are for folks in terms of accessibility. There are folks that have completely different learning styles. There are folks that take in information at different levels in different wavelengths of things. And I say that all to say that, like, you know, I would like to see a world where— and it was scarier with— to keep with the OtterBox or not OtterBox. Sorry, that's not— OtterBox is not a sponsor of this. But the Otter.ai note-taking tool was that after I got an email randomly from the note-taker to all the people, also to all the people that were in that call with a a narrative recap of everything that we talked over. It wasn't a transcript. It was a narrative recap, which is fine enough. Okay. There were— there was a screenshot of just a random person that was on the call that was also there. And also what's most scary for me, I think, or just very concerning, is at the bottom it was like, here's the productivity score of the call, 84%. Here's the engagement of the call, 72%. And it's like, where Where's at least the— at the very least, where's the asterisk at the bottom that says, this is how we calculated this whatever? And I immediately go, I'm not a pessimist, but in that moment I was like, this is going to be used by people in higher positions, people in power to wield over folks in middle management and direct service to say, hey man, you didn't have an 84% or higher engagement score on our last Zoom call. You are now on a personal improvement plan. And that is a scary place to be. And so I think less about like, these tools are what they are, but I think about the people and the systems and the toxic systems at times that sometimes wield these brand new shiny tools in a way that doesn't feel good and also is working against their mission and against their employees.

Kate MartignettiIt's time for Tony's Take Two.

Tony MartignettiThank you, Kate. And thank you, for supporting Nonprofit Radio. I'd like to say thanks every once in a while because I don't want you to think that we're taking you for granted. I'm grateful, grateful for your listening. And if you get the insider alerts each week, I'm grateful that you get those letting us into your inbox. This week I'm in Portland, Oregon recording a whole bunch of Good, savvy, smart interviewers for upcoming episodes. Hopefully that helps like show our gratitude because we're out here collecting good interviews for you to listen to if you can't make the nonprofit technology conference yourself. So thank you. I'm grateful that you listen. Grateful that you're with us week after week. That's Tony's Take Two. Kate?

Kate MartignettiThank you guys so much for listening to us every week. We appreciate you. Well, we've got beaucoup but loads more time. Let's return to Artificial Intelligence for Nonprofits Redux with Justin Spelhaug, Amy Sample Ward, and Tristan Penn.

Amy Sample WardTristan, I love that you brought that up. Don't love that it happened. Right. But you brought it up as an example here for folks because I think it's an easy entryway into a conversation on one of the points Tony mentioned at the start of the call, like what are some of these preconditions? And you were like, oh, people are like, oh, bright shiny, right? That's what we do. Oh, bright shiny. Like I'm gonna use this tool that like took the notes in here and A place where we've seen for many, many years in NTEN's research is that nonprofits struggle. This isn't to say that for-profit companies don't also struggle with this, but nonprofit organizations struggle with consent. They struggle with privacy and security.

Tristan PennMm-hmm.

Amy Sample WardAnd so here's a well-meaning, well-intentioned, right? I'm gonna use this tool, except it's emailing you. You didn't consent to that. It emailed all the participants in that call. There was no opt-in, right? Let alone a very clear opt-out. Like, why did I even And get this. That's not even to say opting into sentiment analysis of whatever is a community Zoom call, right?

Tristan PennRight.

Amy Sample WardAnd so, when we peel that back and say, okay, well, we just wouldn't use that note-taking app, right? Sure. But when we're thinking about preconditions for this effective work as an organization, do— what are your data policies in general? The number of organizations that we work with that still don't have a data policy because they think, Well, isn't there like some law about data? So like, we— why would we have our own policy? Okay, there is some law related to data, right? Different types of data have different laws, but that's not the same as an organization saying, what data do we collect? Why do we collect it? How long do we retain it? What if somebody wants us to remove it? How do we do that in our systems, right? So this level of, uh, fidelity to your own data to your own community members, to the policies that you've set up to manage those relationships and trust for so many organizations are already not in place. Or like I said, there's just not a fidelity to them that makes them trusted. Right. So then to say, oh yeah, we're ready to add this note-taking app to our community calls or our client calls, it just— that's the place where I have the most fear is actually not the tools having bias. I know they have bias, and that is a place of concern and a place we can, can address it. But my— the most fear I have is people still operating within that without any of the structures or policies or training to deal with both maybe bias in a tool they use and their own bias or their own issues, right? And it accelerates the harm that can be created in that.

Tony MartignettiAmy, I want to use some of that to, to go to Justin, uh, about something very closely related. Uh, the, the, uh, the nonprofit leadership role, the responsibility of, of nonprofit leaders. I think it gets to a lot of what Amy was just talking about, but what do you, what do you see as the, the responsibility of nonprofit leadership in, in formulating these policies, but also in just You know, making sure that the preconditions are there so that we, we can be successful in integrating artificial intelligence, whether we're bringing in an exterior, an outside tool, or, or, or building our own even. And that, that may be a, that may be a big lift for a lot of listeners, but, but generally the, the nonprofit leadership's role.

Justin SpelhaugYeah, I mean, there's a lot of a nonprofit leadership plate today, and I think we have to meet, uh, leaders where, where they're at. And, and I think the Very first step, and Amy mentioned this in the very beginning of the call, is raising the capacity of their knowledge and of their staff's knowledge of how these tools work and what are the edges of the tools and how to apply them effectively in the flow of work. And there's training available. As an example, we have a 4-hour course on LinkedIn. You don't need to do it all at once, but it's actually pretty good. It's for— it's not for developers, it's not for techies, it's for frontline program staff, fundraising staff, fundraising staff, finance staff, the ED, to really learn about how to think about these tools. With that knowledge, then you can take the next step, which is starting to engage in, I think, simple ways to apply these tools to get on-the-ground experience of what they're good at and what they're not good at. Using things like— I'm from Microsoft, so I'll mention Bing or Microsoft Copilot. To look at writing donor appeal letters or whatever the process may be. They can just start learning about these fundamental language models and what they're good at. I think it's important as an organization thinks about getting deeper into AI and really thinking about how do they apply it to their processes, whether that be fundraising, whether that be engagement with beneficiaries, that they think really deeply about data and data classification and that, that, that It gets a little sophisticated, but just ensuring that we've got a strategy to use AI for the data that we want to use AI for, and that we segment data that we do not want AI to reason on away. So start with getting the basic skills built out. A lot of organizations I meet with are just at the very beginning stage of that. Use the simplest tools to accommodate the job to get some experience, and then start to think longer range around data, data classification, and more advanced scenarios that can be applied.

Tristan PennTony, can I—

Tony MartignettiJustin, what's that 4-hour course on— Tristan, just let me drill down on a free resource. I love free resources for our listeners.

Justin SpelhaugYeah, I'll pull it up while Tristan answers. It's a LinkedIn course, Nonprofit AI Fundamentals, but let me get that for you here.

Tony MartignettiOkay, Tristan, go ahead.

Tristan PennYeah, can I— I really like how Justin initially said, you know, there's a lot on nonprofit leaders' plates already too. In terms of responsibility, and I want to gently push back and answer your invite to call you in, Tony, in the premise of the question, which was, what is the responsibility of nonprofit leaders now? And I would say, yes, obviously there's a responsibility, as Justin has illustrated, that like we need to be better in terms of strategy, in terms of tech, in terms of AI in general on how we fold these crucial tools in. But I would also say that there's an equal and almost larger responsibility on those who fund nonprofits. I think a lot of times in the nonprofits that I've worked with, interacted with, and worked within, their operational and financial model has been very ham-handedly built in a very Dr. Seussian way, which doesn't really make sense at times. And it's because year after year there are different grants, different fundings that require different things at different times based on whatever the hot new term is. 10, 15 years ago, it was mentoring. So a lot of times everything was geared towards mentoring. And I say that because this implies that a lot of these nonprofits are already built on a structure that is very shaky. And so there's a lot of other things that need to be done. But I do think a big responsibility sits with folks who fund nonprofits, foundations, and also local governments, federal— the federal government in making sure that when they are pushing a grant or putting out an RFP for a grant that says you need to fold in tech and you need to fold in AI in this way to get kids to learn or get kids in seats in the classroom, that you're doing so in a way that creates longevity and solid solid nonprofit organization operational work and just doesn't like slap an iPad in front of a kid. And I think that's really— I used to work at Boys and Girls Club, so that's where I always default. But I think that I've, based on my experience, it's always been a really weird way of having— going into a financial model of an organization year after year because it's like, oh, well, we started doing that because last year's grant for it. So now we just do it into perpetuity. And so again, you have that weird Dr. Seuss-style way of thinking. And I think funders and grant folks can do a lot by being very clear and forward-thinking in how they are offering up these monies.

Kate MartignettiIt's time for a break. Virtuous is a software company committed to helping nonprofits grow generosity. Virtuous believes that generosity has the power to create profound change in the world and in the heart of the giver. It's their mission to move the needle on global generosity by helping nonprofits better connect with and inspire their givers. Responsive fundraising puts the donor at the center of fundraising and grows giving through personalized donor journeys that respond to the needs of each individual. Virtuous is the only responsive nonprofit CRM designed to help you build deeper relationships with every donor at scale. Virtuous gives you the nonprofit CRM, fundraising, volunteer, marketing, and automation tools you need to create responsive experiences that build trust and grow impact. virtuous.org. Now back to Artificial intelligence for nonprofits redux.

Amy Sample WardYou know, that's not only the mindset like this, this— it feels like they're being strategic by saying, oh yeah, well, we were able to come, we were able to pitch that in a way that we got the fund, but then that's changing their strategies all the time. It also, back to the point before, is meaning the data you have to work with inside your organization is, okay, well, 2 years we structured it this way, for 2 years we structured it this way, Do we even have like a unique ID to connect these people and say, oh, they were in both of those programs? Like our own datasets are messy and influenced by funders saying, oh, now we need you to collect these demographic markers. You know, and it's, it's, we, we as organizations are often pressured by those funders to do it the way they want because it's easier for them and tells the story they want to tell. But that's really, really messing up the data sets and the program kind of, uh, processes or, or business processes that we have in place. And I, I just wanted to connect that to broader things that NTEN's worked on and advocated for for many years from the Equity Guide specific to funders, and that is that funding technology projects takes time and it takes a lot more money than like $30,000 for whatever the licenses are for something, right? Like, it's not uncommon that an organization building a model, an internal use model— this isn't some big flashy commercial thing, this is just for them to, you know, like I said before, identify program participants that maybe, you know, could use intervention— it's not uncommon that would take 2 dozen tries to get the right model in place, right? To really make sure the algorithm is is fine-tuned, that the outputs are appropriate. Well, you can't go through 2 dozen models in, in 3 months, right? And then have something there. A nonprofit would need a couple years. And our, our funders, there's already plenty of funders saying like, ooh, now we have this AI grant, you know, opportunity.

Tristan PennMm-hmm.

Amy Sample WardAre, is that grant gonna be comprehensive of the work to get their data in a good place, to get their program staff ready and trained? To Justin's point, every staff person really trained adequately. on, on not just what are these tools, but what's a good prompt, what's a good use case for this, right? All of those pieces so that they can adequately and materially contribute to then what is this project we wanna do, what is the best fit for us, and how do we, how do we build it?

Justin SpelhaugAnd, and just to add on, and we'll wrap up, uh, to Amy's point and Tristan's point, AI hasn't changed the fundamental physics of what makes a good technology project. I mean, it's It's people, it's process, it's tools, it's capacity building, it's a long-term strategy. All that is the same. And if your listeners are wondering, where do I even get started in understanding the language of this stuff? Because you asked the question, it's called Career Essentials in Generative AI. It's on LinkedIn, it's free, and I've taken it. It's pretty good. So I think it's worthwhile for your listeners.

Tony MartignettiThank you, Justin. How about NTEN, Amy? Resources for folks? I mean, hopefully they're already going to the Nonprofit Technology Conference where there are going to be a lot of— there are a lot of sessions on artificial intelligence. I know because I'm going to be interviewing a bunch of those folks. So this is probably the second of, I don't know, 6 or 7 AI episodes in different— around different subjects. Yeah. But NTEN as a resource for learning, Amy?

Amy Sample WardWe have lots of them. There's, um, you know, work— uh, not workbook, but like a guide. There's of course the equity guide. There's some materials on the website. We have an AI course and other courses that talk about AI. There's community groups where you can ask questions, and of course the conference. But, uh, thanks to Microsoft and Okta who gave us some supporting funding. NTEN, along with Institute for the Future and Project Evident, are at the tail end of a community design process where we've worked with over 40 organizations in this process to create an AI framework for organizations, whether you're a nonprofit or not, who are trying to make decisions around AI. And our framing for this is the framework for an equitable world. So it isn't just that you are a 501 registered in the US, right? Or that you're a grassroots organization in whatever country. Like, if you want to live in that equitable world, then this is the framework that we can all share and work in together. We're going to do a little preview at the NTC and have whoever comes to the session, is going to get to road test it with us, and then we'll publish it publicly after the NTC. So lots more, and obviously I'll share that with you when it comes out. But what's really, I think, important from this is that it is a framework that is built on the idea that all of us are part of these decisions, that all of us have responsibility in these decisions. Yeah. And that all of us are accountable to building that equitable world, right? This isn't, you know, the quote unquote responsible tech, or this isn't like, this isn't just for those projects where you're, where you're gonna do something good over here. This is whatever we're doing, it's gotta be good. It's gotta be building us into an equitable world because what else are we doing here, right? If it's not for that. And so I'm excited for folks to get to use it. It'll be published. Yeah. For free everywhere. Anybody use it, please go, you know. So lots more on that soon.

Tony MartignettiAmy, you are perfectly consistent with the framed quote that you have behind you. All of us are in this life together. You're living your framed art. I admire it. Justin, we've got maybe 10 minutes left. What would you like to talk about we haven't touched on yet? Or go further on something we have?

Justin SpelhaugWell, no, maybe, maybe I'll just build a little bit on what Amy was— the question you asked of what are the resources available. I think that's pretty useful to the organization. So one is, one is the training that I mentioned. Two is we just recently ran a Nonprofit Leaders Summit where we had 5,600 people together, about 4,500 online, about 1,000 in a room talking about how do we grapple with AI? How do— what are the use cases that make the makes sense. How do we think about data security and privacy? And we're going to continue to invest in that. We're going to be rolling that out more globally as well with events in Australia and others. But that convening and that dialogue and just getting the community in dialogue, I think, is so important. I learned a ton from that. We're also going to continue to push on affordability and making sure that we've got affordable access to our technology so that every organization can use things like Microsoft Copilot for free, providing, you know, they've got access to our nonprofit offers. And then finally, innovation. And I'm interested in looking at scenarios that span the sector where if we invest once, we can create a multiplier effect. And one of the areas that we're partnering on is with Save the Children, Oxfam, and many other organizations on the Humanitarian Data Exchange, which is a large dataset used to help organizations coordinate humanitarian and disaster relief domestically and internationally in a more effective manner so our missions don't overlap. But that dataset hasn't been super useful to date. Applying things like language models, training on that, and creating a toolset that is cross-sector for many organizations, you'll see us continuing to invest in that way. And I look forward to ideas from our NTEN partners here on the phone as well as you know, the community at large on where we can make bets that will really help the sector together move forward.

Tony MartignettiTristan, what would you like to touch on or go deeper in? We've got 7, 8 minutes or so.

Tristan PennYou know, I think I just want to underscore what Amy was talking about and that we've all been working on, which is the— All right,

Tony MartignettiI'm a little tired of you underscoring Amy, Amy underscores you, we force each other. You know, I agree with that. You should have seen us. We work together. It's getting a little dull. It's a little dull now. All right. No,

Tristan Pennyou should have seen us when we were in office. Our desks were 20 feet away from each other, and there was a constant— there was a worn line in between our desks,

Tony Martignettiright? And nobody wants to be in between in that 20-foot, uh, in that 20-foot space. All right.

Tristan PennUm, I, I will say, um, being a part of, um, the community group, what Amy was saying, um, about working with 40 other organizations to figure out what a healthy and robust and equitable process is for any organization to interact with and field AI is crucial. And I'm so glad that we are able to be a part of it and we're going to be debuting it at NTC. It's something that I've learned a lot from just based on someone who, again, like I said before, I came from Youth development. My degree is in child psych. So, but I've learned a lot over the years working with NTEN, working at NTEN. But I think one thing that's, that's been really, really beneficial is learning from all those folks in the group. And a couple of things that did come up in when we were creating that framework, which was that organizations are making all kinds of decisions Every day today, and I will say that it kind of highlights that we are talking about AI and how it like will look, sound, and feel, and how it looks. This is all kind of we're not meaning it to be, but it's all within a vacuum, and we can't think like that. We can't think all of us who have now we are four years out from 2020. Our lives were forever changed, and every nonprofit will have their own sad story to tell about how the the pandemic impacted them. And I say that to say is that like no no. No one was prepared for that. And so if we keep on talking about or playing around with this idea of AI as like it's going to solve problems or it's going to sit in this world in this vacuum, we're not doing ourselves justice and we're being very forgetful about the past that we just went through. And so if we're able to instead consider how AI will interact with the dynamic world that we all live within, that's going to better behoove us both individually but also organizationally when we're planning strategically. Yeah. If that's year after year for you, if that's every 5 years, I don't know what that is. So having that strong tech baseline for folks. And then I think also the other thing is people in all roles are considering AI and aren't sure how it applies to them. I think staff, we've read stories that AI will replace workers, but have no idea what to do with, you know, where, where that fear sits with them too. It should just add to their work and not replace them. And I think a lot of we're seeing, you know, I'm— I am on TikTok. And so, you know, that's a whole other like bag of algorithms and like, you know, things that we can dissect and pull apart. But I do— there are a lot of stories of, you know, there are folks getting laid off left and right. And I would have to, you know, that begs the question why generally, but also like what is the role of AI in all of this too? I think it's really interesting when layoffs happen at a time when AI is accelerating in a lot of our worlds. Whether it's in tech and whether it's in other sectors across the world. And I think that there is a lot to be done by organizations who don't fall prey to like the siren song of like AI and are going into it clear-minded and not saying, oh, well, we can cut out this department and put it in, put, you know, this learning module in or this, you know, I think that's, that's really where you're going to see a lot of organizations and organizations and companies thrive as opposed to just laying folks off. A lot there.

Tony MartignettiNo, we're— yeah, we're taking it in. Yeah, no, and, and the reality is that I admire the, the consistency between you and Amy, uh, and, and, and NTEN generally. I mean, I made fun of you, but what it shows is you're all thinking the same way, you know, you've all got this, the same concerns for the nonprofit community.

Tristan PennHuman first, human first, you know, like We're all humans and we're all prioritizing us as humans. And if we start prioritizing other things, then it's not going to go well.

Tony MartignettiWell, but at NTEN, you're walking the talk, so, and consistently. Amy, you want to take us out with all of us are in this life together?

Amy Sample WardYeah, I mean, I think the biggest thing I want folks to leave with is that That future is not predetermined. We are not sitting down and saying, well, okay, like, I'll wait for my assigned robot to come tell me what to do, right? It is still up for all of us to write that every day. And the people who most need to have their sentence at the start of the article or whatever, you know, at the start of the book, are the folks who are being told in a lot of different systemic media type ways that they do not get to have their sentence in the article, you know. And so I, I hope that nonprofits know this is both an opportunity to shape and influence as AI tools are being developed, to shape and influence the tools that we build within our sector for ourselves with our communities. But it's also a responsibility for nonprofits who are the ones often closest to and most trusted by those systemically marginalized communities who are experiencing the most real-time harm to be the supporter that brings them into that work. They are not necessarily going to get tapped by a company to learn this or do whatever, even though I hear Justin saying these, these, you know, opportunities are free and accessible. You as a nonprofit can say, we think we might build something. Can you be in our design committee? Can you work with us? We'll make sure that we all learn together, right? As an organization they're already in relationship with, they trust, they've, you know, maybe benefited from programs or services. You have the responsibility and incredible opportunity to be the conduit for so many communities to enter this, this quote unquote AI world. And that's a really important, I think, gift, you know, that we have as a sector to, to be the ones helping make sure so much, so much more of the world is part of developing these tools and designing them to be accountable to us as people.

Tony MartignettiThere, Amy Sample Ward, our technology contributor here at Nonprofit Radio and the CEO of NTEN. Also Tristan Penn, Equity and Accountability Director at NTEN, and Justin Spelhaug, new Corporate Vice President and Global Head at Technology for Social Impact at Microsoft. My thanks to each of you. Thank you very much. Real pleasure.

Justin SpelhaugThank you.

Amy Sample WardThanks so much, Tony. Thanks, Justin.

Justin SpelhaugCheers. Thanks, guys.

Amy Sample WardJustin, I'll see you in

Justin Spelhaug20 feet.

Tristan PennNo. Thanks so much, Tony.

Kate MartignettiNext week, the generational divide.

Tony MartignettiNow, this is interesting because we've been promising this for a couple of weeks now. And it hasn't materialized. It's very relieving to have someone, an associate producer, who I can blame for this show having been promised— the Generational Divide having been promised for weeks on end and not coming through. Even though it doesn't matter that the associate producer Kate has nothing to do with booking the guests, that the host takes care of that himself, that's irrelevant. I blame the associate producer. And this, this show, The Generational Divide, had better come through next week or there's going to be a shakeup.

Kate MartignettiI'm the one who just reads the script, Tony.

Tony MartignettiOh yeah, minimize the— okay, your title is not script reader, it's associate producer.

Kate MartignettiWell, if you missed any part of this week's show,

Tony MartignettiI beseech you. Look, I was slow on my cue there. I beseech you, find it at tonymartignetti.com.

Kate MartignettiWe're sponsored by DonorBox. Outdated donation forms blocking your supporters' generosity? DonorBox— fast, flexible, and friendly fundraising forms for your nonprofit. donorbox.org. And by Virtuous. Virtuous gives you the nonprofit CRM, fundraising, volunteer, and marketing tools you need to create more responsive donor experiences and grow giving. virtuous.org. Our creative producer is Claire Meyerhoff. I'm your associate producer, Kate Martignetti. The show's social media is by Susan Chavez. Marc Silverman is our web guy, and this music is by Scott

Tony MartignettiThank you for that affirmation, Scotty. Be with us next week for Nonprofit Radio, big nonprofit ideas for the other 95%. Go out and be great. Thank you.

Nonprofit Radio