top of page

AI Is for the Dreamers, Not the Experts

Updated: Mar 14, 2023

Artificial intelligence (AI) is a very good aid to the imagination, just not good enough to imagine for us.

By Malaika Cheney-Coker

Artificial intelligence (AI) is unimaginative about its own utility for the social sector. For example, when prompted on how it can be used within nonprofits to foster creativity, the now famous chatbot ChatGPT dutifully returned answers that boiled down to using AI’s analytic capacity – analyzing large data sets, better understanding target audiences – rather than its generative abilities, like coming up with ideas for solving social dilemmas.

It turns out that ChatGPT was being modest. When prompted, it can round up proven approaches and ideas for everything from beautifying a commercial strip in a way that yields social and environmental dividends, to avoiding common pitfalls in peer mentorship models, to addressing food insecurity in sustainable ways. It can even opine not only on what has been but what could be, e.g., what if there were more coalitions, or more imagination, in the social sector?

For areas in which one has expertise, the results can be disappointing, returning a roundup of what is already known. What AI can do is be an aid to the imagination, and thus to designing programs and initiatives that we would not have arrived at on our own. To avoid being disappointed by predictable results in areas in which we are subject matter experts, we can simply ask AI to show us some non-conventional or experimental approaches to the same topic. I tried that for several things, some relatively arcane (e.g., non-conflict-focused alternatives to narrative structure in fiction) and it did not disappoint.

As an aid to the imagination, AI can take us to the water’s edge of possibility, inspiring us by the daring acts of others, and even giving us cautionary tales (if we request them) of what can go wrong. Interested in Universal Basic Income as an experiment? Ask AI where it has been tried and what its record has been. Ask AI for examples of where a systems lens has successfully applied at large scale to a social problem. Or simply ask it for unorthodox ways of keeping staff engaged. It can speed up the pace of ideation by serving as an encyclopedia-brained member of your team. The tool can also predict some of the inevitable pitfalls and gravitational forces that your ideas will face, as when I asked it what a hypothetical evaluation of Community-led Total Sanitation projects in Uganda would reveal.

When AI does disappoint in its answers – and it will – this is a reminder that AI is only as good as what has been conceived of before. It is a literature review of sorts, of published human thought, and as any good researcher will tell you, new research exists precisely because the existing literature has been reviewed and found wanting. This is where our human superpowers come in – to dream of what has not been done before – or at least not in the particular way we have in mind; to relinquish the sure wins of our expertise in favor of playing the experimentation lottery; to push the boat away from the shore of the known, into unchartered waters.

ChatGPT had this to say about the concept of moving into the unchartered, rendered within the prompt as a need for imagination in the social sector:

Imagination allows us to look beyond what is currently in place, and to envision new possibilities. In the social sector, this means that organizations and leaders can explore new ways of working, new models for delivering services, and new approaches to addressing the root causes of social problems. By embracing imagination, the social sector can find new and more effective ways of creating positive social change.

If you’re thinking, “I could have written that,” you’re right. Perhaps the answers that AI gives us are only part of its magic. The other part may be what happens when we pull out of ourselves the questions to ask of it and the sudden expansion of possibility that results from our curiosity.

29 views0 comments


bottom of page