AI and the social sector

Ah, AI. Can't avoid it.?

I've been to the conferences and workshops, read the listserves, talked to the researchers and read (some) of the research, played with the public tools. The Blueprint 2024 lays out my thoughts on nonprofits, philanthropy and AI for 2024.?

This coming Blueprint (available live and free on 12/15/23) skips the prediction section - and explains why. But I have some thoughts on how AI is going to unfold in the sector, especially after checking out this new resource from Giving Tuesday - the AI Generosity Working Group .

Year 0-1:? November 1, 2022 - 2023 - hype, fear, webinars, and conference talks. Lots of press. Lots of handwaving. Gadgets like this (https://www.nonprofitama.ai/ ). Lots of executive orders and unfunded government mandates and policy proposals pushed by tech companies.

Year 2 - 2024:? More hype, lots of feel-good examples (the Red Cross is using it! AI for disasters!) and a few scandals (lawsuits over data use, data loss, etc) will fill the news. Lots of nonprofits will try things and realize they don't have the expertise on staff, are distracting resources from mission, and will go back to ignoring the topic. By this time, we'll all be using AI all the time, however, as AI capacities will be fully baked into every software product you already have - every microsoft product, Canva, Zoom, Salesforce. We're already there, actually.

As individual organizations become dependent on AI-powered software tools, we'll reach the next level of concern - the existing regulatory regime for nonprofits and foundations will be leaking and breaking, and proposals for new structures and laws will be circulating. The sector's policy advocates will bemoan their missed opportunities, back in 2023 and 2024, to influence the regulations on AI itself. The blend of nonprofits and commercial activity and/or nonprofits and political activity, will complicate such new debates. By this time, the academy and independent research groups like AJL or DAIR will have repeatedly documented harms caused by AI and have proposed numerous remedies.

Having been ignored by industry for 4+ years, we'll see new attention to these ideas. We'll also see a burst of former AI company employees "whistleblowing" or "following their consciences," leaving industry and setting forth to solve the problems they helped create while on the inside. By the time this happens, everyone will be used to and dependent on their AI-enabled tech, and even those who are eager to stop using it will find it "too difficult" to change their tech.

Some of the above is tongue in cheek. But, like the Gartner hypecycle , this loose set of predictions is based on the experience of other breakthrough technologies. It's probably too linear - and doesn't take into account the innumerable "wild card" events that are likely to occur between now and 2028. In other words, by 2028 we'll be having the debates about AI that we had about social media in the 2016 election. Some of these we're already having - especially with regard to elections - and that's a good thing. But it's not going to stop, or even redirect, this flow of events.

It doesn't need to unfold this way at all. Sadly, I don't see enough activities, organizations, advocacy, push back, regulatory oversight out there to prevent this (all too familiar) pattern from playing out. And certainly not compared to the dollars that are being spent now by corporate marketing departments to hook nonprofits.

?