Most books about artificial intelligence want something from you. Some want you excited. Some want you frightened. A surprising number want you to sign up for a course.
My new book, Plainly Speaking: A Candid Conversation About AI for Associations, Nonprofits, and Colleges, wants something much less dramatic. It wants you to make up your own mind, which, it turns out, is a very unfashionable thing for a book about AI to want.
Some context. My first nonprofit job was in Washington, D.C., around 1990. The office ran on WordPerfect, floppy disks, filing cabinets, and a Rolodex, and when you needed to get a document to someone right away, the fax machine felt remarkably close to magic.
Since then, I’ve watched email, the web, cloud storage, smartphones, and video meetings each arrive at work with great fanfare and then quietly become ordinary. Some delivered on their promises. Some mostly moved the work around. AI is the newest arrival, and it’s getting the loudest welcome of all.
The trouble is that most of the people I’ve worked alongside over the years don’t have a team of AI specialists down the hall. They have meetings to run, members, students, and donors to serve, budgets to manage, and at least one thing that was supposed to be finished yesterday.
What they mostly get is a steady stream of headlines telling them AI will either fix everything or ruin everything. Neither headline is much help on a Tuesday afternoon when you’re trying to decide whether it’s okay to paste the board minutes into a chatbot. (It usually isn’t. There’s a whole chapter on that.)
So I wrote the book I wanted someone to hand me.
It starts with the basics: what these tools actually are, why there’s no such thing as a perfect prompt, and why the most useful thing you can give AI is often context, not clever wording. Then it moves into the organization itself, where things get more interesting and considerably less tidy.
Privacy. Time. Disclosure. Policy. Responsibility. AI hiding inside software you already pay for. AI that can build things and increasingly act on your behalf. What these tools really do to people’s workloads. When you owe someone an explanation about how something was made. And how to create some reasonable organizational guardrails without spending fourteen months writing an AI policy.
The back of the book gets practical. There’s a sample one-page staff guide, a policy framework, and a checklist for checking what AI gives you. That last one matters.
I learned the importance of checking AI’s work the hard way. I once asked AI for research to support something I was writing, and it handed me several citations that looked perfectly legitimate. When I went looking for one of the articles, I couldn’t find it. Two of the others were real. One didn’t exist at all.
Nothing about the imaginary one looked any different from the real ones.
That moment shaped this book more than any new feature or product ever could. The question stopped being “What can this thing do?” and became “What happens when we actually put it to work, and who’s responsible when it’s wrong?”
That’s a much less exciting question than whether AI will transform civilization. It’s also considerably more useful when you’re sitting in an office trying to get something done.
If you’ve read What Am I Missing?, you’ll recognize some of the thinking. That book was personal: what talking to AI taught me about my own thinking. This one faces outward, toward the associations, nonprofits, and colleges where so many of us spend our working lives, and the very practical questions they’re facing right now.
The examples in Plainly Speaking are composites, not anyone’s actual office. So if you think you recognize your workplace in one of them, that probably says more about how common these situations are than it does about your coworkers.
A few honest disclosures, since the book would be annoyed with me otherwise.
It doesn’t sell AI, and it doesn’t warn you away from it. In its last chapter, it admits that much of it will age, because anything written about this technology is a snapshot of a landscape that refuses to hold still.
And if you look closely at the cover, you’ll notice that the word Plainly has two letters in a different color. That’s the only trick in the book, and I’ve decided to call it a design choice.
I also didn’t want to end the book with a grand prediction about where AI is taking us. I don’t know.
Neither does anyone else.
What I do know is that people in actual workplaces are already making decisions about AI, often with little time and incomplete information. They’re deciding what their staff can put into these systems, when AI-generated work needs to be disclosed, how much checking is enough, which efficiencies are actually efficiencies, and where the human being still needs to remain firmly in the middle of the process.
I wanted to give those people something more useful than another prediction. I wanted to give them a better conversation.
Plainly Speaking is available now in paperback. If it helps someone have one clearer conversation about AI at work, with a colleague, a team, a board, or even themselves, it will have done what I hoped it would do.
The dedication says it better than I can here: “Use the tools. Explore them. Push them. Have some fun with them. Just don’t stop thinking because something else is willing to do it for you.“