This past week, I attended a packed book tour event at the Crystal Ballroom in Somerville, where I saw one of my favorite sci-fi authors, Becky Chambers, talk about her new book. One of the last questions she took from the audience did not surprise her: What did she think of Artificial Intelligence? She came prepared for this question with a handmade zine titled “ I Do Not Like These Robots”, in which she describes exactly how she feels about her life’s work being “thrown into the great and terrible chicken nugget machine made by some of the worst people on planet Earth.”
I was bombarded with AI talk this week. Two days after the book event, the nonprofit I work for was honored at an event in Cambridge that featured a discussion with MIT professor and Economics Nobel laureate Daron Acemoğlu. Though his new book is about the endangerment of liberal democracy, the conversation veered into AI fairly quickly. The last question from the audience, promised to be “a quick one,” was “How do we talk to our kids about AI?” “That is not a quick question,” I muttered to the colleague beside me.
Though Acemoğlu’s answer was thoughtful, it was by no means satisfying. I don’t believe that anybody can answer that question well right now. Much like the COVID epidemic, our kids are looking to us for guidance on a problem that we Generation X and Millennial parents have never experienced. Unlike COVID, the threat of AI is not going to diminish any time soon. We’re learning with them, and they should know we’re struggling, too.
At my work, we have all been given access to Claude, and we are expected to use it. I quickly learned that when I created a presentation that was a hybrid of my own thinking and Claude’s aggregated “thinking,” my ideas were approved quickly, while the AI’s were sent back for refinement. I now spend hours writing all the content and let the AI do all the time-consuming formatting. It saves time, but it does not save thinking. When I have a good experience or bad experiences using AI, I share them with my kids.
Sometimes it’s just a telling experience. My co-workers and I ran an experiment of sorts with Claude recently, asking it how many “e”s there were in the word “seventeen.” It consistently answered “three.” When we asked it why it thought that, Claude provided the correct answer, explained that this was a common issue for it, and then wrote this: “The fix, basically, is forcing myself to do the boring mechanical thing (spell it out, tag each letter) rather than trusting an intuitive-feeling answer.”
Yikes. I don’t go to machines for intuition. I go to them for the boring mechanical things. I also don’t go to AI to help me write this column, even as I cram to make my deadline. That would defeat the purpose. I write to get these thoughts out of my head, and I can’t outsource that. AI is here to stay, and it will enable some amazing things, but as my daughter writes her college essays, I hope she actually writes them, and that they are not an aggregate of thousands of other essays processed through the “great chicken nugget machine.”
Eric J. Perkins writes about Gen X for The Belmont Voice. When he’s not writing, he’s the Director of Transformation at Addgene, a life sciences nonprofit in Watertown.
