“Keep an eye to the future, an ear to the past, and after thinking it over notice nothing much lasts.” —Robert Hunter, singer-songwriter
Artificial intelligence (AI) has been very much in the news lately. Most of us have heard the warnings about how AI can lead to the destruction of humanity. It is a well-known fact that any advance in technology will lead to new possibilities for abuse. But these advances can be a powerful force for good as well.
For a definition of what AI actually is, I turned to a corporation that has been working with these matters for quite a while: IBM. On its website, it states, “At its simplest form, artificial intelligence is a field, which combines computer science and robust datasets, to enable problem-solving.” It is worth noting that an AI program called Deep Blue, running on an IBM supercomputer, defeated the reigning world chess champion in the 1990s.
You could go back centuries to find warnings about our creations turning against us. Mary Shelley’s Frankenstein or Karel Capek’s R.U.R. come to mind. IBM rightly begins its AI timeline with Alan Turing’s seminal work “Computing Machinery and Intelligence” from 1950. In it, he described the Turing test, wherein a human would send text messages to someone they cannot see. By the responses, they must decide whether the correspondent is a human or a clever machine masquerading as a person.
Musing on AI Capabilities
I suspect the reason AI has become such a phenomenon is that the technology has made great progress in creating sounds and images that could be mistaken for real people and things. This led me to develop a wish list of what AI might accomplish.
Text-to-speech functionality has been with us for decades. In the 1980s, physicist Stephen Hawking brought this into the mainstream, but his robotic voice would not translate well to the reading of audiobooks, unless they had been written by Hawking himself. Now AI can emulate any voice. Imagine the possibilities for ebooks. You could have James Dean reading East of Eden and other John Steinbeck works, Peter O’Toole reading the works of Thomas Hardy, or Cary Grant reading light mysteries. The technology is there, but the legal issues with the estates of those actors would be the real concern.
A few months ago, I met a Long Island fisherman who had fallen off the back of his boat while the rest of his crew was asleep. By the time the Coast Guard had mounted a full-scale search, he had been in the water for hours. As he held on to a flotation device, he could see planes and helicopters flying by and boats with a quarter mile of him. The technology running the search was human eyes behind binoculars. Imagine if the Coast Guard could use high-definition, real-time video matched with an AI program to find any anomalies in the water. The technology exists—somebody just has to implement it.
Speaking With an Expert
I had the good fortune to have a video call with Almin Surani, who holds the title of global nonprofit digital transformation lead at Avanade, an Accenture and Microsoft company. He started by telling me about an elderly woman who was scammed by an AI machine that emulated the voice of her grandson perfectly. He told me that there is so much possibility for evil in technology that those of us who are trying to make the world a better place must work harder than ever to make AI perform for good.
Surani is a Toronto-based IT expert who has worked with nonprofit enterprises for decades. He began his career with the Canadian Red Cross, rising to a 10-year run as CIO from 2008 to 2018. He told me that he spotted the convergence of AI and nonprofit enterprises years ago and has finally landed in a position where he can make those possibilities a reality.
We talked about some of the specific things that AI could do for a nonprofit. Heading the list was finding ways for organizations to do a better job of getting to know their donors and tailoring communications to the individuals instead of just sending out blanket messages. We discussed possibilities such as symphony orchestras sending personalized streaming playlists to their major donors.
I mentioned that AI is shaping up to be a problem in education, as we are already seeing sites advertising “homework helpers”—AI-based programs that can write or rewrite student essays. In the old days, teachers who suspected a problem could simply search the internet for certain strings of words and root out plagiarism in a hurry. There is no such simple solution for an essay improved with AI. However, Surani told me there’s already a program on the market that sniffs out AI work in student writing. To prove that this is a slippery slope, I searched and found at least one program that claims to bypass AI detection checkers. There seems to be no end to this. Surani said that at all times, the IT professional has to be aware of ethical issues in improving the technology.
Writing on the Avanade Insights blog, Surani states that “the sudden popularity of ChatGPT is a testament to its potential to revolutionize the way we interact with technology. With its ability to generate human-like responses, it’s poised to have a significant impact on many industries, including the nonprofit sector.” He told me during our conversation that AI programs like ChatGPT can improve communications with donors in nonprofit organizations by being on the job 24 hours day.
What Others Have Said About AI for Nonprofits
I found a number of sources covering the issue of AI for nonprofit organizations. In June, Lisa Stiffler wrote “How Are Nonprofits Using AI and ChatGPT? The Focus Is on the Donor Dollars” for GeekWire. Similar to what Surani said, Stiffler writes that the most important thing here is donor relations—finding new likely donors and improving communications with the ones you already have. She reports that Microsoft is developing a new suite of software to serve the nonprofit community.
In “Artificial Intelligence: Why the Nonprofit Sector Should Pay Attention,” an April blog post for Independent Sector, Manuel Gomez cautions against using AI to cut costs by replacing people with machines. If handled correctly, AI improvements can follow past technological breakthroughs by taking on routine tasks, thereby freeing up staff to do more effective work. Gomez describes how one university official asked ChatGPT to describe the benefits of AI to nonprofits. I tried that question on Google’s Bard service and got this response:
Artificial intelligence (AI) has the potential to revolutionize the nonprofit sector by helping organizations to:
Automate repetitive tasks … This can free up staff time to focus on more strategic activities.
Analyze data. AI can be used to analyze large amounts of data to identify trends, patterns, and insights that would be difficult to spot manually.
Make more informed decisions. AI can be used to provide nonprofits with data-driven insights that can help them to make better decisions about everything from program development to resource allocation.
Personalize communication. AI can be used to personalize communication with donors, volunteers, and other stakeholders.
Develop new products and services. AI can be used to develop new products and services that can help nonprofits to achieve their goals. For example, AI-powered chatbots can be used to provide customer service, AI-powered predictive analytics can be used to identify high-risk clients, and AI-powered fundraising platforms can be used to automate fundraising.
AI for Libraries
As someone who usually covers the library automation beat, I was curious about how libraries are using these new AI tools. In reading Nick Tanzi’s “How Can AI Be Used in Libraries?” for The Digital Librarian, I got the impression that lots of libraries are looking at the possibilities of AI, but very few have developed practical realities for it. Tanzi opines that it would most likely be appearing in libraries by way of database, streaming service, and ILS vendors.
When I began my professional library career in 1990, we were excited to be on the cusp of an information revolution—possibly the greatest since the invention of the printing press. Now, just 3 decades later, we are doing it again. These are heady times.
When I looked over the literature for AI for nonprofits, there was so much consistent reporting that I wondered if some of it was written by AI itself. That inspired me to try a quick experiment: I copied the article that you are now reading and pasted it into an AI writing program. The free version of Wordtune could only do an abstract rather than a full rewrite. Except for one howler—“A Long Island fisherman fell off the back of his boat while the rest of his crew was asleep. The Coast Guard searched for him using human eyes behind binoculars.”—it worked pretty well. You can read the results at terryballard.org/isetl.html. I was curious to see if the AI version of me writes better than I do, but not curious enough to buy the enhanced program.