A woman is reportedly ending her 12-year marriage after an interaction with an AI chatbot, ChatGPT, convinced her that her husband had been unfaithful.
The unnamed mother of two, who resides in Greece, turned to the AI tool not for advice or conversation, but for a rather unconventional purpose—fortune-telling.
According to reports from the Greek media, cited by the Daily Mail, the woman asked the chatbot to perform a reading based on the coffee grounds left at the bottom of her husband’s cup.
This unusual method, known as tasseography or tasseomancy, is traditionally practiced by fortune tellers who interpret shapes and patterns formed by tea leaves or coffee remnants to predict the future.
It is not a practice associated with artificial intelligence tools like ChatGPT.
Nonetheless, the woman uploaded a photo of her husband’s coffee cup into the chatbot and requested a “reading.”
To her shock, the AI allegedly indicated that her husband was either thinking about or already involved in an affair with a woman whose name begins with the letter “E.”
The chatbot even suggested that the supposed other woman had intentions of destroying their marriage.
This unexpected response reportedly triggered a major marital crisis.
The woman’s husband later appeared on the Greek morning show To Proino, where he shared his disbelief over the incident.
He explained that his wife had previously been influenced by supernatural beliefs, including a past experience with an astrologer that took her a year to overcome.
However, he noted that the AI’s prediction seemed to have a much stronger impact.
“I thought it was nonsense and laughed it off, but she took it seriously,” the man recounted. “She asked me to leave, told our kids we were getting a divorce, and shortly after, I was contacted by a lawyer. That’s when I realized she was serious.”
Though the husband initially resisted the divorce, he was officially served the papers just three days later.
His legal team is now contesting the divorce proceedings, arguing that the AI-generated reading holds no legal credibility and cannot be used as evidence of wrongdoing.
The strange case has drawn attention online, particularly on Reddit, where users responded with a mix of amusement and concern.
Some joked that AI might be putting psychics out of business, while others warned about the risks of relying too heavily on language models for serious decisions.
One commenter described an experience in which ChatGPT inaccurately insisted on basic facts, using it as an example of how flawed AI output can be.
Another expressed deeper concern, suggesting that tools like ChatGPT may blur the line between reality and fiction for vulnerable individuals.
“This kind of technology is powerful, but without the right safeguards, it can seriously distort someone’s perception of reality,” one person wrote.
“Unfortunately, the moment to put those safeguards in place might already be long gone.”