How Nonprofits Can Tap Technology in Lean Times and Avoid AI Hype

Jim Fruchterman helped found one of Silicon Valley’s first AI startups in the 1980s. At Calera Recognition Systems, he and his team created a machine that could scan and read text aloud.

Fruchterman saw potential to adapt the technology into a reading device for the blind, but the company’s board rejected the idea because the market was too small to be profitable. Within months, Fruchterman left to launch Benetech, a nonprofit that went on to incubate dozens of “tech for good” projects, like a library of accessible digital books and a secure database for human rights workers.

Through Benetech — and later Tech Matters, which develops open-source tech tools for nonprofits — Fruchterman has spent more than three decades proving technology can drive social change. He’s also earned accolades along the way, like a MacArthur “genius” grant.

In his new book, Technology for Good: How Nonprofit Leaders Are Using Software and Data to Solve Our Most Pressing Social Problems, Fruchterman offers a practical road map based on his life’s work.

He spoke with the Chronicle about how nonprofits can fund technology and build capacity in a period of cutbacks, how AI grant making is changing, and what gives him hope about the future.

This interview was edited for brevity and clarity.

You’ve described the nonprofit sector as being stuck in a “social sector time machine” when it comes to technology. What’s the cost of that delay?

Software and data are there to make people more powerful and to give them more knowledge and insights. If you’re not very good at technology, it means your team is less effective and you know less about how effective your programs are.

You’ve seen a lot of promising and not-so-promising projects. What’s one common mistake you still see even smart nonprofits making?

The most common mistake is building technology for the way you think people should work or what they should want instead of actually finding out what they need and want.

Many nonprofits hesitate to invest in tech because of donor pressure to limit overhead. How do you think the sector needs to rethink that conversation?

Some of it is financial jujitsu with funders. Instead of saying that technology is part of overhead, you say it’s part of programs. You have to know how you’re finding your new clients, how much it costs you to acquire them, how much it costs you to serve them, and did you make an impact in the end. Every donor wants to know if their grant’s having an impact. Well, that seems like it’s pretty program-related as opposed to this other thing, overhead, which must be minimized.

There’s also a structural challenge in how we fund stuff. The nonprofit sector is dominated by projects over long-term products or enterprises. One-year grants are not the best thing to sustain building a long-term enterprise that’s going to change the system.

Many nonprofit organizations spend less than 3 percent of their budgets on technology, according to the Chronicle’s recent survey. What does sustained, effective funding for nonprofit tech look like in practice?

Let’s get tech spending to 5 percent. That’s less than half of what an equivalent business would spend on tech, and that just hobbles organizations.

Any tips for bringing in that investment, especially as the nonprofit world is reshaped by cutbacks in government funding?

The field is going through some wrenching changes right now. Nonprofits are trying to figure out how much money they’re going to have this year, and philanthropy seems to be drawing back.

Technology is probably one of the few ways that you can do more with less and maybe deliver something close to the quality of what you were doing before but with fewer staff. To do that, you have to actually train your people and invest in technology tools, which is doubly hard when unrestricted support is in short supply and donors really don’t want to fund the core technology.

I tell people, put a proposal together that’s 80 percent about strengthening your data infrastructure and your team capabilities, and then put a 10 percent or 20 percent AI cherry on top. You can’t possibly do AI without doing that 80 percent, and that’s how you actually can get money during a fad.

We’ve seen many foundations in the past year or so announce new AI-focused grant making. How is philanthropy’s approach to funding AI for the social sector changing?

AI grant making is getting segmented based on where the donor is coming from.

A fair number of tech donors are worried about existential threats, which I find kind of overblown right now. There are certainly donors who are working on more humane AI and more human-centered technology. Others are pushing for AI regulation. The industry’s mainly successful in killing that off, but some of it’s getting through. There’s a segment of donors that is focused on AI as a societal threat to child safety online.

Then you’ve got the donors who are seeing this through the lens of their portfolio priorities, like NextLadder Ventures, which will invest $1 billion for tools for frontline workers in areas the funders care about. It’s one of the better ideas that has come out of philanthropy in a long while.

AI tools are developing rapidly. What’s your top advice for nonprofits that want to integrate AI into their work?

My number one piece of advice is wait until there’s a product. It’s going to be cheaper. It’s going to have guardrails and safety. It’s going to have training. There are going to be consultants who actually know how to apply that product to your needs. And you’ll have peers you can call up for advice.

How is your own team at Tech Matters using AI?

We’re big fans of using it to speed up data entry. One of our products is for national child helplines. We have helplines that are now serving 40 percent more kids with the same staff. It’s not the miracle of firing all your staff and replacing them with the AI chatbot, which has not played out well.

Throughout my organization, people are using it effectively and safely to code better. We’ve already deployed summarization tools to squeeze a 400-word answer into a 250-word grant application box.

What advice would you give a nonprofit leader with limited resources and no CTO who wants to start using technology more strategically?

Get a volunteer board member who is a strategic tech leader. Not to do the tech work but to avoid you getting swindled by a consultant or someone selling you a product that’s based on snake oil. You just need someone who has that nose for whether something looks like it could work and looks like it’s at a reasonable price point. I’ll put in a plug for Board.Dev, a nonprofit that’s trying to place more tech leaders on nonprofits boards.

The other thing is to talk to your peers. The odds that there’s someone in your field that you trust who has figured out how to equip their team with better technology is really high. If you can’t find anyone, that’s a signal. Maybe don’t be the first guinea pig in this area if you can’t afford tech and you don’t have the internal tech skills.

You’ve been at this intersection of tech and social good for decades. What gives you the most hope right now?

We’re getting to an era where we’re going to be using tech a lot more and a lot more effectively. Often big crises are a turning point when people start to change. Whole subsectors of the nonprofit sector have been blown up and are gonna have to get reconstructed. Software and data are gonna play some kind of role in seeing that happen. It’s the combination of a number of crises plus the popular awareness of generative AI. People are thinking about it now. And I’m actually quite optimistic that the timing is good.