Over the past few months I’ve received four separate emails offering to create hundreds of genuine (that is to say AI-generated) reviews of books that I’ve published, including one book that’s now over thirty years old.
On top of that, an email offered to create an AI Me, a large language model (LLM) trained on the books I’ve written. AI Me would converse with readers, answer their questions, even sign their copies of my books (digitally, of course), leaving me free to write more or to go into stasis, which is becoming an increasingly attractive option right now.
First, though, were the emails offering to generate thousands of reviews of my work. In one email, the leader of a supposed book club promised to grow my “visibility through honest reviews, promotion, and Goodreads Listopia specialist.” Maybe the awkward phrasing is supposed to show that the message was written by a real person, but I think it’s just AI.
Here’s more of this bot’s AI slop: “The way you connect language, identity, and social change showing how pronouns have shaped rights, debates, and everyday interactions is both fascinating and deeply relevant.” Not even my mother could write a blurb like that.
Another book-review offer that managed to evade my spam folder came from a “person” (surely a bot) who claims to control “a private community of 3,000+ active readers and reviewers who engage seriously with nonfiction especially works on language, culture, policy, history, and social power.” Do I want to hire a private literary militia? I don’t think so.
According to that bot, my book “carries a message that deserves to reach more hearts, and I’d love to help make that happen through a thoughtful, genuine review” (the emphasis on genuine is a sure sign the reviews are fake). After praising me for my insight on a “subject that sparks arguments at dinner tables, classrooms, and courtrooms alike,” there’s this twist of the knife: “…and 4 Amazon reviews. 😶📉” OK, that bit about only four Amazon reviews does sound like something my mother would have said, only without the emojis.
Here’s the bot’s final pitch: “Do you want this book to remain a well-argued whisper in a very loud room…or would you like it placed in front of readers who are ready to engage, respond, and amplify the conversation it was written to start? 🔊📚” I mean, really, what choice do I have?
To no one’s surprise, those writers who choose to pay for each genuine review will wind up with no reviews and no refunds. The private militia will vanish into cyberspace, with maybe one last message invoking its right to keep and bear arms.
I wasn’t tempted to pay $20 a pop for fake reviews – I wouldn't pay for real ones either. But would I pay an entrepreneur (that is, another scammer) to create a fake AI Me? That’s right, with no effort on my part and for only pennies a day, I could be digitally reincarnated.
The service offering to clone AI Me uses for its name a skewed spelling of ‘shrine,’ which suggests both the divinity and the death of the author, neither of which appeals to me. This shrine-bot promises readers that they can “talk to an AI built around a real thinker.” Yes, that’s right, AI is now a noun. If I sign on, I could become that real (holographic) thinker. My response to this offer? “Help me Obi wan kenobe.”
A digitally animated, fully-articulated, talking AI Me is the latest installment of a story that goes back at least to Frankenstein, if not a whole lot earlier. And the thing that's created is always a monster.
The 1953 movie “Donovan’s Brain” offered an AI Me for the analog age. Megalomaniacal millionaire W. H. Donovan dies in a plane crash near the remote mountain lab of a scientist who’s been keeping monkey brains alive in fish tanks. Seeing an opportunity to take his research to the next level, and with no Human Subjects Board to answer to, the scientist harvests Donovan’s brain and places it in a brain-sized tank filled with electrolytes and other nutrients. There the evil brain continues to grow, and surprise, surprise, Donovan’s brain begins to control the scientist, who suddenly starts smoking Donovan’s cigars, walking with Donovan’s limp, and driving into the city to cash Donovan’s checks.
The only way to stop Donovan’s brain is to destroy it, but the brain, intent on world domination, fights back. Finally, in a brief moment away from the brain’s influence, the scientist connects a lightning rod to the fish tank, because a thunderstorm is going to hit the mountain. You can imagine the rest. The movie is on the tubes, if you want to watch it.
Anyway, Donovan’s Brain is what I immediately thought of when I got the AI Me email. Would a cigar-chomping AI Me try to take over the world? Would my analog brain find itself in a digital fish tank generating evil code? Would lightning strike twice and save my readers from AI Me and all the other real thinkers and the megalomaniacal billionaires who want to upload their consciousness to the web?
Yes, I know that all new communication technologies generate fear along with enthusiasm (I wrote a whole book about that, which real me will be happy to discuss with you). And I know that there are good uses for AI. But just to be safe from the dangers of AI Me, I’ve dragged that email to the trash, emptied it, and to make doubly sure it was stake-through-the-heart dead, I rebooted my computer.
Or did I?
