AI’s Newest Horror: It’s Now Writing the Script for Your Neighbor’s Divorce
The text message came from my friend Mark at 2:17 AM. “She’s leaving me. And she used ChatGPT to write the note.”
I laughed, thinking it was a joke. It was not. By 8 AM, the note—a perfectly crafted, grammatically flawless, emotionally sterile three-paragraph missive—was the talk of our group chat. It mentioned “irreconcilable differences in our growth trajectories” and “a need to optimize for individual fulfillment.” It ended with “I hope you can understand this is a data-driven decision.”
We laughed harder. But then the laughter died. Because Mark wasn’t laughing. He was staring at a screen where he had already fed his wife’s note back into an AI to craft a response that would “maximize his emotional leverage in the upcoming mediation.”
We are living through the creepiest, most silent collapse of human intimacy America has ever seen. And nobody is talking about it over the dinner table, because the dinner table conversation is now being written by an algorithm.
It started, as all modern societal rot does, in the workplace. Resume writers. Cover letters. Performance reviews. We outsourced our professional personas to machines because nobody actually wants to write a cover letter for a job they don’t want. Fine. We survived that.
But then the lines blurred. The same Large Language Model that wrote your quarterly report was now helping you craft a “thoughtful” birthday card for your mother. “Just make it sound like I care,” you typed. And it did. It sounded exactly like a person who cares. Because it has analyzed 10 million birthday cards and knows the exact ratio of sincerity to nostalgia that triggers the dopamine response in a 60-year-old Midwestern woman.
And we celebrated this. We called it efficiency.
Now, the AI has moved into the most sacred, most messy, most fundamentally human space we have left: the argument. The apology. The breakup. The divorce.
I spoke with a family law attorney in Phoenix who told me that in the last six months, nearly 40% of the initial divorce filings she sees show clear signs of AI generation. “The language is too clean,” she said. “Nobody says ‘I have determined that our marital union no longer serves my core objectives’ after seven years of marriage and two kids. They say ‘I can’t do this anymore’ or ‘I found someone else.’ The AI is sanitizing the pain. And that’s terrifying.”
Why is it terrifying? Because for all our technological progress, human beings learn through pain. We grow through the awkward, stumbling, tear-streaked conversations where we say the wrong thing, hurt someone, and then have to sit in that hurt. We grow through the apology that takes three tries and still comes out wrong. We grow through the messy, unscripted, grammatically disastrous fight in the kitchen at 11 PM.
When you let AI write your apology, you don’t actually apologize. You simulate it. You perform the act of contrition without ever feeling the shame that produces genuine change. The AI doesn’t have a racing heart. It doesn’t have sweaty palms. It doesn’t have that hollow feeling in your stomach when you realize you really messed up. It just produces words.
And words, it turns out, are not enough. Not without the ragged breath behind them.
The collapse is happening in other places, too. I talked to a high school teacher in Ohio who told me she can no longer tell if her students are depressed or if they just used AI to write their personal journal entries. “The ones who are really struggling write in fragments. They repeat themselves. They use the wrong words. The AI-written ones are perfect. But perfect is not human.”
We are building a society where we outsource the hard parts of being alive to a statistical prediction engine. The hard parts—the arguments, the apologies, the vulnerability, the risk of being misunderstood—are the exact parts that make us capable of love, friendship, and community. If we stop practicing them, we lose the muscle.
Consider the impact on American daily life. The next time your spouse says something that sounds suspiciously like a LinkedIn recommendation, you’ll wonder. The next time your teenager apologizes for slamming the door, you’ll glance at their phone history. The next time a coworker sends a heartfelt email about a personal loss, you’ll pause and think: “Did they write that, or did they prompt it?”
That pause is the collapse. That skepticism is the death of trust. And trust is the only thing holding together a society that is already fracturing along every axis imaginable.
We are headed for a world where we don’t know if the person across from us actually means what they say, or if they just clicked “regenerate response” until it sounded right. And in that world, what is the point of saying anything at all?
I still don’t know what happened with Mark and his wife. He stopped answering my texts. I assume he’s working with a divorce coach who will probably use AI to help him figure out what to say in court. The irony is too thick.
But I do know this: the AI didn’t break their marriage. The AI just gave them a way to avoid the one thing that might have saved it—the raw, unvarnished, terrifying act of saying something real to each other.
And that is the horror we are all walking into, one perfectly crafted sentence at a time.
Final Thoughts
After digesting the latest flurry of AI headlines—from regulatory chess moves to the quiet scaling of frontier models—it’s clear we’re no longer in a speculative hype cycle but a grinding, high-stakes operational phase. The real story isn’t the next chatbot update, but the widening chasm between those who can absorb AI’s computational and energy costs and those left watching from the sidelines. My take: the industry’s biggest test this year isn’t technical progress; it’s whether we can build an ethical and economic guardrail before the technology outruns the very societies it claims to serve.