Over the past 2 years, I’ve read several hundred articles about AI, specifically AI in writing because as a writing teacher, I felt that my world ended with the release of ChatGPT in late 2022. While I no longer lose sleep every night as I did 2 years ago, I still brood for hours on end as I continually look for answers: How do I teach with (or without) this technology? Can I still produce authentic writers and critical thinkers? Will anyone put hard work into writing ever again? 

The most difficult part is that there are no right answers. The technology changes weekly, my curriculum changes continuously, and I no longer feel sure of anything. For a type A person like me, that’s unsettling, maddening, infuriating—and sometimes (maybe?), a little exciting.  

Ethan Mollick’s book is a decent overview of AI—especially for people who have read little and want to better understand its capabilities. But for educators, particularly those who have been immersed in this space, and even more so for writing teachers who have been hit the hardest by this technology, the book doesn’t offer many specifics. In fact, Mollick’s tone feels dismissive (and frankly, a bit pompous) toward those with specific concerns, and he often contradicts himself.

Early in the book, he posits that the legal implications related to training LLMs using copyrighted information is unclear (since AI doesn’t store text directly; rather “weights” or “patterns” from its training in which tokens/words are likely to follow) and that it may “not be ethical as AI companies do not ask permission from the people whose data they train on” (34). Later he writes “I won’t delegate my writing to AI because of the delicate issue of copyrights and the law.” And still later, he offers several examples of how he used AI help to write the book (from summaries of articles to paragraph revision ideas), but he does not credit AI as a co-author or as co-intelligence, saying “because AI is not a person but a tool, I will not be thanking any of the LLMs that played a role in the creation of this book any more than I would thank Microsoft Word” (214). So apparently it’s okay—highly encouraged even— for everyone else to use AI to write, but not him, and when he uses it anyway, he sees no need to cite it. Sorry, but that wounds my writing brain and heart. Always cite your sources, especially those that give you specific words and ideas.

Two of his mantras are “Always invite AI to the Table,” but (to be sure AI is giving the best output), ensure there is always a “human in the loop.” Okay, that makes sense. Especially for his Wharton MBA students, some of the brightest in the country and in their mid 20s or early 30s with years of critical thinking skills, thus the ability to work effectively with AI. Later he does concede that we might become too dependent on AI (51), that it should be as “assistive tool, not a crutch” because if AI writes all our first drafts, we could “lose our creativity and originality” (119) along with the “quality and depth of our thinking”…relying on AI to do the “hard work of analysis and synthesis” (120). 

He never discusses the reality that students will (of course) use it as a crutch and ditch the hard work. When does AI come to this table? 4th grade? 9th? Many of our high school students would gladly offload all of their thinking to AI. Later still, he says that “our AI future requires human experts…facts, reading, writing, history, and all the other basic skills so that we continue to have educated citizens rather than delegate all our thinking to machines” (191). So how do our students become experts at anything if AI is always invited to the table? These are questions he does not address. 

This is where he comes off as naive: Use AI for everything because it will be so much better!  Make sure you’re already an expert in the basics so you can use it well! Thankfully, an AI Writing group that I follow clarified that Mollick is often “out of his lane.” He knows how effective and exciting AI can be in his MBA entrepreneurship class, but he generalizes that to all of education, which is ludicrous and insulting to K-12 educators who are trying to navigate this tricky landscape of teaching basic skills and critical thinking while tech companies are selling us on AI integration into everything we do. 

Yes, the outcome with AI may be stronger and more efficient, creating “impressive essays that exceeded their initial attempts by refining and redirecting the AI” (60), but is that how learning works? Writing, for one, is not efficient, nor it is meant to be because it involves thinking and analyzing (which Mollick admits to right after he says he’s never delegated his own writing to AI), and the epiphanies along the way that may emerge slowly. An AI writing coach could be hugely beneficial (Khanmingo seems to be on the right track), but the AI available today is far more than a coach: it opens the door for students to do as little as they want.

Finally, I’m so tired of the calculator analogy: “AI brings fear just like the calculator once did, and now we know that the calculator is just a tool to solve more complex problems.” True, but the calculator does not complete a 3 page math problem, including the written explanation, with a single command and the push of a button. It’s a stupid analogy.

I am glad I read this book. It offers an overview, presenting possibilities, ideas, and uses of AI, including some exciting ones and lots of career related ones. But as another member of my writing group mentioned, we know too little about AI and its best use and human brains and development for Mollick to have written this book this soon. He did not have enough time to fully process implications, heed his own advice, understand the need for some guidelines, or see his own blatant contradictions. He just wanted to be the first one to pen a book (with AI’s help and no citations/acknowledgments—oh yeah, I already mentioned that).

Read the book. Let me know what you think. AI is here—like it or not.