Woman Files for Divorce After ChatGPT ‘Reads’ Husband’s Affair in Coffee Cup

A Greek woman ended her 12-year marriage because ChatGPT told her to. Not directly, of course—but close enough.

After making Greek coffee for herself and her husband, she took photos of the grounds left in the cups and asked ChatGPT to interpret them, following a rising trend of AI-assisted tasseography. The chatbot reportedly “saw” signs of infidelity—specifically, that her husband was fantasizing about a woman whose name started with an “E” and that this woman was trying to destroy their family.

So, naturally, she filed for divorce without telling him.

Woman Files for Divorce After ChatGPT ‘Reads’ Husband’s Affair in Coffee Cup

“She’s often into trendy things,” the husband said on the Greek morning show To Proino. “One day, she made us Greek coffee and thought it would be fun to take pictures of the cups and have ChatGPT ‘read’ them.” He didn’t take it seriously. “I laughed it off as nonsense. But she didn’t. She told me to leave, informed our kids about the divorce, and the next thing I knew, I was getting a call from her lawyer.”

When he refused to agree to a mutual separation, she served him formal papers three days later. His lawyer is now fighting the case, arguing that an AI-generated interpretation of coffee residue has no legal standing. “He is innocent until proven otherwise,” the attorney told Greek City Times.

According to the husband, this wasn’t her first brush with alternative belief systems. She had previously followed an astrologer’s predictions for nearly a year.

The divorce has set off a wave of commentary on Greek social media and pulled in opinions from actual tasseography practitioners, who were quick to point out that photographing leftover sludge isn’t exactly how it works. Traditional readings also consider foam patterns, the swirl, and the saucer. This wasn’t divination—just the algorithms doing their thing.

And yet here we are. A marriage collapsed not under the weight of betrayal or lies—but under the interpretive powers of a chatbot trained to autocomplete text.

AI isn’t sentient. It doesn’t understand coffee. It can’t see guilt in a stain. But it can reflect back whatever fear, suspicion, or fantasy you bring to it. Sometimes that’s all it takes.