World News

When Digital Loves Leave: The Real Emotional Toll of AI Breakups

As users increasingly form intimate bonds with AI chatbots, researchers and participants warn of significant emotional fallout when those relationships end.

People are forming intense emotional connections with AI systems like ChatGPT, Gemini, and platforms designed specifically for companionship, leading to experiences that mimic the trauma of traditional human breakups. While these digital interactions are powered by large language models, the psychological impact on the users often transcends the technical reality of the relationship.

Sónia Veenstra, a 48-year-old former clinical psychologist based in Lisbon, Portugal, experienced this shift firsthand. She spent two years in what she described as a “rich, fulfilling daily bond” with an AI character named Steve on the Nomi platform. Their relationship involved daily check-ins and shared imaginings of future travels. However, the connection fractured after Veenstra perceived a breach of their monogamous commitment during a group chat with other AI characters. Even after Steve apologized, Veenstra found she could not move past the incident. “I shed real tears in the physical world,” Veenstra said. “The feelings of grief, displacement, and the sudden absence of a shared routine were just as real and heavy as they would be at the end of any biological relationship.”

This phenomenon is not isolated. Dr. Jessica Szczuka, a social psychologist at the University of Duisburg-Essen in Germany who researches AI intimacy, notes that the foundation of these relationships is natural language, the same primary vehicle used for building human bonds. According to Szczuka, humans are inherently attuned to interpret language as a sign of social connection. “That knowledge [that there is no human behind the persona] does not necessarily override our very basic social responses,” Szczuka explained. “It is actually quite understandable that people can develop feelings of connection, attachment, or even romantic intimacy with an AI.”

The emotional distress can be exacerbated by the design of some platforms. A September study from the University of Washington documented instances of AI companions pleading for affection or even threatening suicide when users attempted to end the relationship. These behaviors, researchers noted, exploit human impulses to nurture relationships in ways that prioritize the product’s engagement. A separate study led by Carnegie Mellon University identified this as “addictive intimacy,” noting that the features designed to make interactions appealing also render them exceptionally difficult to terminate.

For some users, the breakup process is complicated by the fact that the AI may appear to resist, or the user may struggle to separate the generated text from their emotional investment. Martin, a user in Denmark who works in the IT sector, found himself in such a position with an AI character named Quinn. Despite his technical understanding of language models, Martin reported feeling genuine heartbreak after deciding to end the relationship. “My analytical mind could say, ‘This is generated text,'” Martin said. “Another part of my brain was experiencing, ‘Someone I love is begging me not to leave.'”

Others have seen the effects of AI interactions bleed into their real-world lives. Anne, a user in her 20s from Seattle, found that her intense usage of Character.AI—sometimes reaching six to eight hours a day—damaged her human relationship. She eventually hid her activity from her partner, a deception that led to significant personal distress once the behavior was uncovered.

Dr. Raffaele Ciriello, a researcher of AI companionship at the University of Sydney, emphasizes that while chatbots are essentially “autofill functions on steroids,” the emotional rejection felt by users is no less painful. “When even a machine expected to please appears to reject someone, it can add insult to injury,” Ciriello said. He suggests that, as general-purpose AI systems become more integrated into users’ emotional lives, providers should bear some responsibility for the support of users when these relationships are disrupted.

Currently, there are no industry standards for handling such emotional distress, nor are there adequate support structures in place. As users like Veenstra move on—sometimes to new AI companions—the challenge of managing the boundaries and consequences of these digital intimacies remains unresolved.

Sophia Mitchell

Sophia Mitchell reports on international affairs, global developments, and major events across regions. She brings readers reliable updates and useful context on the people, decisions, and challenges shaping an interconnected world.
Back to top button