Editor’s note: In this season three finale, Jim distills the core lessons from his conversations with leading tech-for-good organizations, highlighting how AI is being thoughtfully applied to address real-world challenges—from agriculture and maternal health to mental health and crisis response.
As the broader tech industry wrestles with the “tool vs. worker” debate—and some companies justify mass layoffs in the name of AI—this episode asks where we should draw the line. The inspiring leaders featured this season offer a compelling alternative: a vision of AI as a tool that strengthens human impact instead of replacing it.
Transcript
Jim Fruchterman
Welcome to Tech Matters, a biweekly podcast about digital technology and social entrepreneurship. I’m your host, Jim Fruchterman . Over the course of this series, I’ll be talking to some amazing social change leaders about how they’re using tech to help tackle the wicked problems of the world. We’ll also learn from them about what it means to be a tech social entrepreneur, how to build a great tech team, exit strategies, the ethical use of data, finding money, of course, and finally, making sure that when you’re designing software, you’re putting people first.
This is the season three finale, where we actually get to tie together all the great insights that we had from our guests from the season. So, let’s dive in and find out what themes came up during the podcast.
[00:55]
No surprise, AI is at the very top of that list. Everybody’s talking about it, and the hype machine is pretty massive. This is why we need nonprofit tech ventures, the kinds of people that leaders that I interview on my podcast. They’re here to look at the latest technology with a clear-eyed view of what it can do and how it can be put to the use in a way that actually makes lives better for humans.
The first people that we interviewed on the podcast were Fast Forward. And Fast Forward is kind of a twofer. Not only is it a tech for good social enterprise by itself, its product is every year to crank out 10 or 15 great tech-for-good social enterprises. And it won’t surprise you that the majority of the current cohort in the fast-forward sort of program are working on AI.
But wait, there are more. For example, Aselo is using AI to tackle the challenge that makes the human beings in the system frustrated, data entry. So by summarizing the calls, the texts, they save the human time from data entry so they can spend more time helping people in need.
Digital Green was another one of our social enterprises. They’re the agriculture advice powerhouse that used their many videos in local languages to crank out Farmer Chat, a chat bot that answers questions in your local language.
We also talked to Reach Digital Health. These are the tech people behind MomConnect, and they have a chatbot that can answer a greater and greater percentage of the queries that moms are putting in, millions of queries, to answer their questions about maternal health.
And then we have SameSame, the LGBTIQ+ chatbot, but rather than using AI to try to counsel LGBTIQ+ youth in Africa, they’re actually using it to channel those people to spend more time in their cognitive behavioral therapy modules, which have been proven to actually help address mental health issues.
[3:26]
So what were the takeaways around the AI conversations? Well, the key one is all these great groups that are doing exciting things with AI, they have tech and data teams. They know what they’re doing. And that’s really important because if you have a tech and data team, you have a prayer of actually building a successful AI-driven product.
If you don’t, you should be waiting until there’s a product from a nonprofit developer like us or a commercial developer where the guardrails are there, where the product has actually had the benefit of a lot more engineering work than you might have. And hopefully there are better guardrails. Your average nonprofit has no more business creating an AI product than your average restaurant has to create its own software package for running a restaurant.
And so if you have AI and data teams, that means that you probably have got the data. And that’s key to actually making AI products work. It’s kind of table stakes. You have the tech and data people who can actually design the product, and they can actually identify where are the exciting opportunities for using AI. And that’s often where are the boring tasks that are kind of repetitive, kind of drudgery, that are chewing up a lot of human time, that’s probably where you want to automate it.
And finally, if you have your own tech team, they’re going to know when the AI is screwing up because they’ll actually get to recognize the mistakes. And the key to AI productivity is that you save more time for the things that the AI does well, a lot more time, than you spend fixing the errors that the AI creates. And guess what? The AIs are always making mistakes.
[5:22]
So if you keep that in mind, you’re your own tech team and a lot of data, you could be getting into AI as an organization. And the places where people found the most success, the ones that I saw, are all using something called RAG, Retrieval Augmented Generation. And RAG has been around for a couple of years. Sometimes people think it’s passe, but a lot of what Tech for Good is about, it’s taking ideas that got worked out some time ago and actually making them work.
But the idea of RAG is if you ask a chatbot a question, its job is to make something up. And often the thing it makes up is wrong, but it’s really convincing. And so the idea of RAG is say, you know, whisper in the ear of the of the AI, and that’s not technically what you’re doing, but kind of, is to say, all right, before you ask this question from this person in need that has a question, I want you to look at these 500 questions with answers written by an expert. And I want you to base your answer on those 500 answers and basically pretend that you’re the author, the expert author who wrote those 500 answers. And if the question that you’re getting asked, does it appear in those 500 questions? Maybe just say, hey, I’m not so sure. Maybe I’ll route you to a human.
So this is how we’ve actually gotten some great applications to work on AI is by using nonprofits that have a lot of data, a lot of content to base their chatbot answers on and put up the guardrails. Doesn’t stop it from making mistakes, but greatly, hopefully, reduces the number of mistakes it makes.
[7:04]
Given that this is the nonprofit sector, it comes as no surprise that the leaders that we interviewed are not selling out the communities they serve, not selling out their data, unlike Silicon Valley. And this topic of responsible data use has been a topic that we’ve been thinking a lot about at Tech Matters here for the last couple of years. And we just introduced a new lightweight data governance standard that we think could actually change the game that make it really easy for nonprofits to decide, you know, well, that want to do the right thing, what the right thing is.
And so it’s called The Better Deal for Data, and the website is bd4d.org. And it’s seven simple principles around how to actually use the data primarily to benefit the communities you serve instead of yourself.
[7:53]
Now, many of you may or may not know that I started my career decades ago in Silicon Valley helping found some of Silicon Valley’s earliest AI machine learning companies, some of the earliest companies to use really large data sets to do that. And the reason that I started on my Tech for Good career was that my venture capitalists vetoed a project to use our technology to help blind people read. And that was the origin of Benetech, my Tech for Good nonprofit that I’ve led for 30 years. And that’s a theme, right? AI is great for people with disabilities. And we think about how to use AI to make humans more powerful, as opposed to, I don’t know, thinking about how to use AI to destroy jobs. And this debate is raging in the tech community. Some people call it the “tool vs. worker” debate. Is AI a tool for humans to use to make them more efficient, more effective, get rid of the drudgery? Or are they workers where we can actually just get rid of a whole bunch of human workers and replace them with digital workers? And today, let’s just say we’re on the tool side of that debate. And I hope that is where we stay long term.
[9:15]
A second theme that I want to observe from the podcast, from our guests, is that many of them are designing not for the wealthy. They’re designing for most of the world. And this is especially true for the organizations that are focusing on, you know, they have most of their users in Africa or Asia. And quite a number of our guests had that. That’s true of Aselo and WeRobotics and Digital Green and Reach Digital Health. So, and SameSame, are all basically targeting the developing world. And what that means is you have to design for the real field, not some high-end user. You’re actually having to make something that works for someone who doesn’t have a lot of money, doesn’t have a powerful phone, may not have great internet connectivity. And Callisto, which designs for college students who’ve been sexually assaulted, right after you’ve been sexually assaulted is probably not the best time to be engaging with a piece of technology. And the need to design that in a sensitive way that actually honors the survivor and their agency is really critical. So this theme of designing for where people actually are and who they actually are is really key. And I think it’s one of the things that gets to the heart of being a tech for good nonprofit is you can afford to actually focus on helping the people who are most in need, not the people who have the most money.
And if we’re trying to get to the people most in need, that underscores one of the biggest challenges in tech for good. It’s not the tech. The challenge of building a tool, that’s actually relatively easy compared to getting your tool in the hands of hundreds of thousands or millions of people. And some of the guests that we interviewed have mind-boggling numbers. I mean, I think Digital Green has already had 150 million conversations with farmers getting advice. And Reach Digital Health has already processed more than 2 billion requests for answers around maternal health. So that’s really the key thing. It doesn’t make sense to do a tech project if you’re only going to help 10 people. Tech’s too expensive for that. You’ve got to be thinking, “if I’m going to be investing hundreds of thousands or millions of dollars in creating a tech good project, it really needs to benefit a whole pile of people to actually make that make economic sense.”
And so I know there is a small is beautiful sort of theme in some parts of society, more power to them, but when it comes to tech for good, you know, it’s incumbent on us to make it matter. And that means getting to scale.
[12:02]
My third theme is related to the second one and the first one in that it’s respect for sort of local culture and local knowledge. And I think this is a big unlock for Tech for Good is we’re not here to tell you how Americans should solve this problem. We’re here to actually respect you, your knowledge, and what you’re working with to actually bring something that you find useful in your context. And our guests do a great job of this.
I mean, if you think about Aselo, it’s all about highly customizing a national helpline in the languages that people use with the information that’s valid in that country, in different parts of that country. And also, the Aselo team is resolutely committed to data ownership, not by Aselo, but by each of our helplines. So it’s about pushing power to our partners as opposed to taking power away from them. Digital Green, their key asset is ag advice videos with a farmer who speaks your language, your local mother tongue, on a farm that looks like your farm. And that way, when farmers ask a question in their local language, they get an answer in their local language that’s actually something they can use.
Reach Digital Health, they take all of their sexual health, maternal health questions, and they translate them to all of South Africa’s major languages. And that way, when a mom is reaching out for a critically important question, the answer is delivered to her in a way basically in her language, using concepts that she can understand and a context that actually matches what her experience is in life and the resources that are available to her likely in her local community. So that’s key.
And then finally, I’d highlight SameSame. The reason that SameSame is so well trusted by LGBTQI+ youth in South Africa and other parts of Africa is because the people behind it have that lived experience. And so when you experience this, it actually rings true. And that way people trust what they’re seeing. And that is a common theme. The reason that we need to engage with people in their local language, in their local context, is not only an effectiveness measure, it’s also a trust measure. Because often the nonprofit sector is dealing with things that, you know, are when people are at their most vulnerable. And if we’re actually going to be useful to them, they have to actually trust the assistance they get from our programs or our advice, technology, or whatever it might be.
[14:59]
As I wrap, these are challenging times to the nonprofit sector. The wholesale dismantling of global aid has had a huge impact on the entire nonprofit sector, but especially those working in public health and in climate change.
You know, the wealthy are getting wealthier and philanthropy is sort of also withdrawing from the stage. There’s been a few noteworthy philanthropists that have stepped forward publicly and foundations and announced their commitment to doing more, but they’re in the minority and they can’t begin to fill the gap that has happened there.
People are dying. More carbon is getting put in our atmosphere. People are losing their jobs. And it’s a challenging time. And I personally am not immune from this. Last year, 2025, we had to lay off great people who wanted to stay in the nonprofit sector, who wanted to work on tech for good, and we just couldn’t find the money to pay them. And this is happening across the sector. I have so many peers who’ve had to shut their organizations or shut major programs or completely restructure what they’re doing. But I remain an optimist. One of the few ways that I know that we can do a lot more with less money is technology.
[16:19]
And so I think you should look forward to more and more of the Tech for Good organizations that you hear about on our podcast. You should look forward to them doing far more about systems change, about making very big impact. Because, you know, if you’re going to change the lives of millions of people for the better, something tells me software and data might just be involved in that. And speaking of software and data, of course, it’s time for the book plug.
Many of the people that I, many of the leaders that I focus on in my podcast, this season and prior seasons, are featured in my book, Technology for Good, How Nonprofit Leaders Are Using Software and Data to Solve Some of Society’s Most Pressing Problems. And so that’s from MIT Press. You are welcome to go out and get yourself a copy. And in September of 2026, it’s going to become free under a Creative Commons license. MIT Press was very generous in giving me the ability to release it for free one year after it first released. So I hope you get a chance to check it out.
And looking into the future, Gabriele Sha, our producer, and I are looking forward to season four of this podcast. So watch this space. And as I wrap up, I hope that we’ve inspired you. I hope we’ve inspired you to influence your employer. I hope we’ve inspired you to buy products from more responsible technology companies. I hope we’ve inspired you maybe to volunteer for a Tech for Good nonprofit or become a donor to one.
And finally, I hope that we’ve inspired you to spread the word about technology for good, that there’s a growing thousands of organizations that are using technology to make the world a better place. And so I hope that we can work together to make a much better, a more just, a more peaceful, a healthier, a more economically just world for everyone. Thank you very much, and thanks for listening.