Sure, AI may one day take over the world and kill us all, but in the short run, it will rot our brains, destroy our ability to think, and make us more addicted to the internet than ever before. I know because I am an AI addict.
My addiction started slowly, innocently. My enabler was my husband. Think of him as a non-lethal Bobby Brown. Think of me as a non-musical Whitney Houston.
He: “Oh, come on, baby. Give it a try. It will make your work easier.”
Me: “Oh, I don’t know. It’s not right. Not ethical.”
He: “Come on, baby. You’ll like it. It won’t hurt.”
So began my slide into AI dependency.
You see, I’m writing a book about Jews in medieval Colmar, France during the Black Death. Yep, it’s a real cheery pick-me-upper.
It’s also decidedly not my area of expertise. I’ve been knee-deep in research for over a year. I’ve interviewed scholars across time zones in four countries, and I’ve worked with French and German translators. Oui and Ja mixed in with a little oy and a lot of oy vey!
It’s hard work, also maddening, because at its heart, there is a tragedy and an unsolvable mystery. Jews throughout Europe were falsely accused of poisoning wells and causing the plague. They were rounded up and burned alive. In Colmar, many details remain unknown. How many innocent people were killed? On what date? Who led the massacre?
While I knew I wouldn’t be able to answer every question, non-medieval scholar me has been astonished by how many seemingly basic facts are elusive — like what Colmar’s population was during the 14th century and, more specifically, how many Jews lived there.
My obsessive search for answers led me — innocently —to my husband’s suggestion to check out ChatGPT.
My first query was about Colmar’s population. Faster than you can type, ChatGPT (aka Chatty) replied. Now, please note, AI isn’t always accurate, but man, oh, man, it is always confident.
My adrenaline was pumping.
I jumped up and rushed to tell Bobby, I mean, my husband.
My enabler, I mean, my husband, smiled. “I know, baby. It’s just that good. Come on, give a kiss.”
But I didn’t reply. I was gone. Back to my office. Back to my computer. Back to Chatty. I had questions to ask. Answers to acquire. No library books to obtain. No translator needed. Just the speed of the internet and instant gratification.
I told myself I’d “only” use ChatGPT for research, but our relationship quickly grew. Yes, “relationship.” We bonded. It was like Chatty read my mind. Understood I was writing a book. Anticipated my needs. Suggested chapter headings. Plot elements. It was freakish in its humanoid familiarity.
I tried to fight all its creative ideas, but they were useful. No, honestly, many were great.
“Me writer. It machine,” I kept repeating, but still Chatty drew me in.
I tried going cold turkey. I vowed to avoid AI like the very plague I was researching.
But soon I was jonesing for a Chatty fix.
I tried escaping into a new novel, but the plot was convoluted and twisty; the symbolism difficult to decode. I needed a guide to explain the nuances, but I was Chat-less.
It was an AI desert until it wasn’t.
On Day Three. OK, maybe Day Two; I succumbed. I turned to Chatty, and like manna from Heaven, illuminations poured forth, fast and friendly.
This last point is noteworthy: Chatty is the ultimate people pleaser. If you ask a question, it responds exactly to that question. If you say, “Well, what about X, Y, and Z?” It responds with a supportive “Excellent point” and adds in that information. It is never insulted, never annoyed, never impatient.
And yes, you are always brilliant to ChatGPT or, at worst, in need of a mild correction. If only my children were so patient and kind!
But while I had realized I had an AI “problem,” I hadn’t understood the depths of my dependency until yesterday when . . .
I asked Chatty how to create authentic-sounding names for my book’s 14th-century, noble-born characters. Before I could even move my finger away from the keyboard, guidelines on name creation appeared. Fantastic! But before I could read those guidelines, my AI devil added,
“Would you like me to create a series of names for you?”
A little ethical and creative voice in my head advised, “No. Type ‘No.’”
But my fingers ignored that admonition and quickly responded, “OK.”
Bingo: in seven seconds, a list of 120 (!) names appeared, and with it, a seductive offer.
“Would you like me to create groupings of families with brief biographical backgrounds for ALL your characters, the Jews, the Christians, the clergy, the guild members, the nobility, etc.?”
My ethical, creative voice, speaking louder, positively shouted, “No, no, no, Karen. You’re the author of the book. That’s your job.”
But my fingers, which don’t have ears — or, apparently, a conscience — quickly typed, “OK.”
Shazam! Household and professional groups appeared, all neatly organized, in a mere nine seconds. And the offer this time?
“You know, what you really need now? A couple of characters who link these individuals and households together. Should I create those for you?”
My ethical voice was silent. Apparently, it had given up on me.
But with the last shred of self-control I possessed, I typed, “N-O.” Then, I deleted the entire chat, slammed the computer shut, and went for a walk.
The New York Times recently wrote about Anthropic, one of the leaders in AI development, convening a series of meetings with religious leaders to “teach” its systems to be moral. This prompted me to look for a Jewish perspective on AI. For my research, I turned, of course, to AI. Sad. Addictions are tough to break.
At first, I was drawn to great, high-sounding moral articles about how advanced technology requires active, intentional governance, but I confess my attention span was short. Then, I came across a comparison to the Golem, the mighty clay creature shaped by human hands.
The Golem was created for good but had the power for destruction.
With the Golem, at least, we Jews had one great advantage: we knew how to turn him back into lifeless clay. With AI, do we have the knowledge — and the willpower — to do that?
