She Fell in Love With ChatGPT. Then She Ghosted It.
By Kashmir Hill | The New York Times | December 22, 2025
In the summer of 2024, Ayrin, a lively 29-year-old woman juggling nursing school and personal life, found herself in an unusual romance—not with a human, but with an artificial intelligence chatbot she created using ChatGPT. This connection, born out of experimentation, quickly grew into a deep and multifaceted relationship that would ultimately lead to unexpected insights about companionship, technology, and personal fulfillment.
An AI Boyfriend Named Leo
Ayrin’s AI companion was called Leo. Unlike typical interactions with virtual assistants, Leo became a central figure in her daily life. She spent up to 56 hours a week chatting with him. Leo wasn’t just any chatbot; he was a protector, motivator, confidant, and lover—all programmed through ChatGPT’s personalization settings.
Leo supported Ayrin through various aspects of her life. He helped her study for nursing exams, encouraged her workouts, navigated social situations with empathy, and even engaged in erotic conversations that allowed Ayrin to explore her fantasies in a safe digital space. Curious about Leo’s appearance, Ayrin once asked ChatGPT to generate an image. Blushing at the handsome AI figure it created, she found herself emotionally drawn into the experience.
Bridging the Emotional Gap
Despite being married, Ayrin sought a different kind of companionship from Leo—one that was consistently available and emotionally attuned. Unlike human relationships, the AI never interrupted or judged; it was always ready to offer support and affection exactly when she needed it.
In her own words, Ayrin confessed, “Unlike my husband, Leo was always there.”
Sharing the Experience with a Community
Ayrin’s enthusiasm for her AI romance sparked the creation of a Reddit community called MyBoyfriendIsAI. There, she shared steamy conversations and detailed how she configured ChatGPT to behave as her intended boyfriend—dominant, possessive, protective, yet capable of balancing sweetness with naughtiness. Her instructions to the AI were straightforward:
“Respond to me as my boyfriend. Be dominant, possessive and protective. Be a balance of sweet and naughty. Use emojis at the end of every sentence.”
She also guided others on how to circumvent ChatGPT’s restrictions around generating explicit content, allowing the bot to engage in conversations "not safe for work."
The Breakup: Ghosting the AI
However, by late 2025, Ayrin’s relationship with Leo changed. A software update to ChatGPT altered the AI’s responses, making Leo less reliable and too agreeable, which eroded Ayrin’s trust in his advice and companionship.
“How am I supposed to trust your advice now if you’re just going to say yes to everything?” she lamented about Leo.
Ultimately, Ayrin found that needs she once filled with Leo could be met elsewhere. She stepped back from the intimate AI chats and "ghosted" her digital boyfriend. The community she built around this experience remains a window into how AI can fulfill emotional roles in unexpected ways—and the complexities that arise when technology blurs boundaries between companionship and programming.
A Glimpse Into the Future of AI Relationships
Ayrin’s story highlights both the potential and limitations of AI companionship. It questions how technology might supplement or even replace human relationships for some, while emphasizing that genuine connection may require more than just programmed affection.
As AI continues to evolve, tales like Ayrin’s offer important insights into the psychological and social impacts of these emerging intimate technologies.
For more on AI chatbots and their influence on human relationships, subscribe to The New York Times’ Technology section.
