“I Am Not a Robot” looks, superficially, like an attempt to assess the value of A.I.—a doomed endeavor, since the technology is always improving. (“One of the biggest obstacles I faced was that the tech kept getting better faster than I could test or write,” she notes.) On a deeper level, though, the book is a performance in which Stern models the process of deciding whether different kinds of A.I. are good for her, as an individual. She notes that she wrote all of “I Am Not a Robot” herself (the words “started in my brain and traveled, via my MacBook keyboard, onto the page”), yet she also employed “BookBots”: custom A.I. agents she built using ChatGPT and Claude. These bots, she explains, had access to her outlines and transcripts, and throughout the writing process they “researched, summarized papers, crunched data, copyedited sections, suggested better words, brainstormed, and even mocked-up illustration ideas.” (When the book was finished, they wrote the blurbs: “ ‘I Am Not a Robot’ is unusually self-aware,” ChatGPT observed.) Stern questions whether using the BookBots was a good idea. Did having them “constantly edit and tighten my writing cost me the version of this book that might have resulted from the slower, more reflective process of figuring out what I actually wanted to say?” She doesn’t really arrive at an answer, perhaps because whatever insight she found wouldn’t necessarily apply to anyone else. A.I. is everywhere, but it’s not a one-size-fits-all technology. It’s something you have to try for yourself, on your own problems, drawing your own conclusions.
Talking to college students over the past few months, I’ve been struck by the terrible bind they’re in when it comes to A.I. On the one hand, it seems obvious that they need to learn how to use the technology, to keep up with the competition and to prepare for the future. On the other, by employing A.I., they may end up cheating both their professors and themselves. Commentators outside the classroom seem to hold extreme views (A.I. is the future; A.I. is wrong), but they aren’t about to enter the job market for the first time. “Should we use it, or not?” some students asked me, recently. I basically said no—but maybe yes, carefully, a little?
Many people are living through versions of this dilemma in their own contexts. At work, it certainly seems wise to acquaint yourself with the tools that are changing your job and your field. But can you do so without losing your skills, and without being accused of faking, cheating, or shirking? In our personal lives, many of us are dependent on smartphones and social media, which we’ve spent decades decrying as oppressive and manipulative. But if we explore A.I.-based alternatives (by using the technology to summarize our e-mails, say) are we engaging in behavior that is in some way antihuman? Some argue for a total rejection of A.I. at work, in art, at school, and at home, while others rush to employ it everywhere. But the views expressed on both sides might not apply to you, in particular, because each of us has different goals, contexts, and competencies.
Stern’s book underscores the insufficiencies of “A.I.” as an umbrella term. When she goes for a mammogram, her doctor shows her how A.I.-based diagnostic tools have already improved her radiology practice: the software, she explains, has meaningfully increased accuracy, helped prioritize the most complex cases, and even boosted morale, by showing overworked radiologists how often they’re correct. And yet, at various dentists’ offices, Stern finds dentists who are “using AI to upsell the crap out of us” by employing tools that claim to identify incipient cavities worthy of early intervention. (“Something like this I wouldn’t even treat,” a more conscientious dentist says of an issue flagged by an A.I. “It’s not worth putting a hole in your tooth to fill just to get a cavity of that size.”) “Technology that reassures in oncology can feel manipulative in dentistry,” Stern writes, because those fields are fundamentally different. Dentistry is rife with judgment calls based on preferences, while oncology isn’t.
