Will the new technology transform the way nonprofits do their work?
Fifty years ago, futuristic fiction offered an even balance between utopia and dystopia. Life in “The Jetsons” seemed pretty great, even if “1984” was not. Today, outside of a small, technophilic Bay Area bubble, it seems that most people expect only zombies when they imagine what the future will bring.
But what if technology were to make things better? More specifically, what if the new kinds of artificial intelligence that we’ve all been endlessly hearing about could safely, ethically, affordably and quickly support the mission-driven work of nonprofits in New York?
What do we mean by ‘artificial intelligence’?
The constant chatter about AI is such that it’s often unclear what “AI” even means — so let’s start with some basic definitions. Artificial Intelligence is any effort to get a mechanical system to act like a person. Traditional AI has been in practical use for years — for example, scanning a check with your banking app or getting Netflix recommendations. Generative AI is the newer breed: tools that have been trained to create entirely new content based on a user’s direction.
What’s even more helpful is to understand the distinction between what I call “exponential AI” and “incremental AI.” Exponential AI gets most of the attention and sounds like science fiction: self-driving cars, new disease cures, chatbots for mental health, etc.
By contrast, incremental AI is the everyday use of widely available, off-the-shelf AI tools in ways that might not revolutionize work but might make it 20% better. Incremental AI helps social workers write bureaucratic letters and notes in half their usual time, freeing them to take on new clients or better serve existing ones. Incremental AI is the equivalent of using Excel or Gmail with basic competency. And just as with Excel or Gmail, truly widespread achievement of basic AI literacy could be transformational for the nonprofit sector — protecting staff, fighting burnout and allowing organizations to serve more people with the same resources.
Right now there is a gap between private- and social-sector takeup of AI. One recent paper finds, for example, that 45% of workers in finance are using AI regularly, compared to only 19% in the social sector. Recognizing this, the AI for Nonprofits Sprint (which I helped to found at the Fund for the City of New York) is designed to help 5,000 staff at 100 New York nonprofits reach basic AI literacy by the end of 2025.
Recently, the AI Sprint surveyed the current state of AI practice for more than 1,200 staff at more than 500 nonprofits across New York State. Fifty-four percent of respondents said that they have used AI at work, and 78% want their employer to offer training opportunities, but only 11% have actually received any guidelines or policies about what is or isn’t permitted. This phenomenon of staff using AI in the shadows means that nonprofits are at risk of inappropriate use. At the same time, they are also failing to benefit from what is essentially free R&D — many staffers may be doing amazing work but they are afraid to share it with their colleagues and supervisors.
In more than a dozen seminars and training sessions over the past six months, the AI Sprint has seen where nonprofits are using AI, how they are using it, what their greatest concerns are and what’s possible going forward. Among other things, nonprofits are using AI to repurpose content. A 40-page academic research report can be plugged into ChatGPT and transformed. You can ask it to write proposals, newsletter content, tweets or webinars. If you are using your own, original information, there are basically no limits.
Today’s generative AI tools are also really good at pretending to be a person. While they’re not as good as the best human role-playing experts, with the paid version, you get an unlimited role-playing partner who’s available at three in the morning and can practice with you 10 times a week. Staff can practice having tough conversations with clients or donors or elected officials as much as they want.
Within nonprofits, the greatest interest and enthusiasm for AI is clearly coming from the executive office and the fundraising department. Executive directors and VPs of development are feeling greater pressure than ever before to do more with less. In many places, executive directors are also dealing with persistent staffing shortages, partly because their staff have told them that too much of their daily work is filled with administrative drudgery.
In many nonprofits, however, the responsibility to explore and execute an AI strategy is handed to the IT department or the general counsel’s office on the assumption that AI is about implementing software or regulatory compliance as opposed to organizational culture and professional development. Unsurprisingly, those nonprofits where less-than-enthusiastic staff are leading the AI exploration have relatively slower processes.
The future of AI: Miracle, menace or meh?
There is a well-documented “hype cycle” with all new technologies, and there are legitimate reasons for concern about what AI might bring to the nonprofit sector. In the short term are issues of bias, accuracy and data privacy. If staff use AI tools for their work, will their work product be of poor quality? Will that work actually hurt those nonprofits aim to serve?
In the medium term, will reliance on AI lead to a loss in key work skills, or even job losses, for nonprofit staff?
In the long term, will AI have a pernicious impact on our natural environment? Or will AI tools lead to a Terminator-like scenario where robots fight against humanity?
Predicting the future is tricky business, but several scenarios are possible.
AI as Miracle (Exponential). AI technology continues to get better, and at a lower and lower cost. New, customized AI tools allow nonprofit staff to do very little boring paperwork and instead focus on human relationships. Low-income New Yorkers apply for and receive government benefits in record numbers. Solutions to formerly intractable environmental and education challenges are found. And AI tools build bridges across political divides, fight conspiracy theories and allow New Yorkers to see their shared humanity.
AI as Miracle (Incremental). AI technology barely improves from what we have today, but most New York nonprofits find ways for their staff to have access to off-the-shelf tools, with clear guidelines on use and some basic how-to support. Nonprofit staff avoid issues around bias and data privacy by never sharing anything they weren’t already sharing anyway through the Microsoft or Google clouds. AI doesn’t do independent work or thinking, but it reduces time spent on first drafts, grant proposals and spreadsheets. From a climate perspective, we learn that AI, while resource intensive, is less so than the video streaming, cellphone use or airplane travel we’ve already been doing. Nonprofit leaders commit to use efficiencies to better serve their missions by supporting more clients, not by cutting staff. And AI drives a revolution in staff equity, allowing individuals from impacted communities to take jobs, communicate across bureaucracy and serve at all levels of our nonprofits.
AI as Menace (Exponential). Nonprofit leaders don’t wait for good, safe new tools to be built — they simply rush headlong into using biased, erroneous, untested solutions. Legal bots give the wrong advice to immigrants and foster care agencies focus on the wrong kids because the computer told them to. AI turns out to be worse for the environment than crypto and cars and cows — and climate change accelerates rapidly. Nonprofits cut many jobs, hurting valuable employees. Those staff that remain rapidly lose the ability (or never gain the ability) to write thoughtfully or reason carefully, because a machine does it for them, albeit badly.
AI as Menace (Incremental). AI doesn’t get much better than it is today, and nonprofit organizations never dive in wholesale — they don’t offer training or clear guidance. But staff do increasingly use it. Despite poor, inaccurate and biased results, they increasingly input confidential data and get bad guidance. And every day that passes, their skills grow worse. AI is no worse for the climate than some other uses of electricity, but that’s of little comfort.
AI as Meh. AI is not as amazing as the tech bros promised, but it’s also not the end of society. It has clear uses for many people and organizations. But the nonprofit sector basically opts out, deciding that it is not worth the time and energy to make sense of a valuable new technology. Years later, nonprofits opt for slightly cheaper, more complicated, less effective versions from nonprofit contractors or get some kind of hand-me-down from a corporate donor.
There’s no fate but what we make for ourselves
The menacing scenarios must obviously be avoided. But so too should the nonprofit sector avoid the trap of a world of “meh.”
In an effort to help nonprofits get past their natural skepticism of the AI hype, and the inertia and bureaucracy that afflict many organizations, the AI Sprint does four things:
- Conducts a baseline survey to give nonprofit leaders clarity on where staff already stand with regard to AI.
- Supports creation of clear guidelines to help staff use AI incrementally and safely.
- Provides basic AI training for all staff in an effort to ensure that this opportunity is not restricted by age, gender or education level.
- Employs a peer-learning strategy to allow staff across organizations to experiment, learn and teach each other.
This is the work of the AI Sprint, but any organization can do it on its own. Even before this work is completed, one thing is clear: The successful implementation of AI is all about organizational mission, entrepreneurial spirit and a relentless focus on staff development. As such, most organizations would be well advised to explore AI out of the executive director’s office, with lots of HR support, rather than the IT department or general counsel’s office.
Fear is not the way forward. The nonprofit sector can do new things. And who wants to live in a world of “meh,” anyway?