The death of a loved one can be devastating, leaving family and friends to cope with overwhelming grief and a profound sense of loss. But some survivors may find comfort and relief in chatting with an artificial intelligence-powered “ghost” of the deceased person, researchers reported in a study published June 12 in the Proceedings of the 2026 Designing Interactive Systems Conference.
The scientists recruited 16 volunteers who were grappling with a loss and open to interacting with an A.I.-generated, text-based recreation of that person.
Before the session, volunteers shared a brief description of the deceased individual, including their relationship, personality traits and communication style, which researchers then provided to the generative ghosts, also known as “griefbots” or “deadbots.” Once the experiment was underway, a researcher reviewed the chatbot’s replies before sharing them with the participant.
Volunteers interacted with two bots that were based on an A.I. model that previously powered ChatGPT. They conversed with each for about 20 minutes during a virtual session conducted over Zoom. One version, which the researchers dubbed the reincarnation bot, chatted in the first person, typing things like, “I remember when we went to the beach together.” The other, known as the representation model, used third person, writing things like, “She loved going to the beach with you.”
The two bots had different effects. “Reincarnation most often created immediacy and closeness, while representation evoked memory and reflection,” the researchers write.
The A.I. ghosts sometimes made up facts about the deceased person, but those inaccuracies didn’t seem to affect participants’ overall experience. They were, however, less enthused by some of the griefbots’ specific word choices, especially terms of endearment that the deceased person never would have used, like “champ” or “sweetie.”
“Maybe the chatbot didn’t know when exactly grandma was born, but if it used some slang or regional word choice that she used, it was extremely impactful,” study co-author Jed Brubaker, an information scientist at the University of Colorado Boulder, says in a statement.
Overall, participants were “largely positive” about the interactions, the scientists write in the paper.
“They love this technology,” study co-author Jack Manning, also an information scientist at CU Boulder, tells Science News’ Kathryn Hulick. “[Participants] clearly came with a goal in mind. They had a question they felt was unanswered or a detail in their life they were desperate to share.”
One volunteer who had lost her mother reported that speaking with the chatbot felt like “getting the closure I needed so bad. … It’s just amazing.”
“I can see her,” the participant said. “I can feel her.”
However, volunteers also understood the potential downsides of chatting with A.I. ghosts. Some worried about becoming emotionally dependent on the bots or experiencing discomfort if the interactions “felt socially or ethically out of place,” the researchers write. “Maybe I use this technology for months, and I start growing more fond of it,” one participant shared. “I don’t know if I would like the person I would become if I kept using this always.”
Others feared that ongoing conversations with a bot might complicate the grieving process—a concern echoed by researchers. “The crucial danger is people don’t need to adjust to a world without this person,” Nora Lindemann, a technology ethicist at Osnabrück University in Germany, told Nature’s Tammy Worth last year. “They can live in a somewhat pretend, in-between stage.”
Did you know? The need for research
When study co-author Manning was 11, he lost his 13-year-old sister to a heart condition. He was initially skeptical of A.I. ghosts. But amid the growing popularity of commercial griefbots, he also understood the pressing need for more research on their effects.
“I felt it was important for me to do the work because the people who are the largest fans might skip the empirical research and just make a product,” he says in the statement.
These issues are exactly why the CU Boulder researchers set out to study A.I. ghosts in the first place. Commercial griefbots from companies like Project December, Re;memory and Séance A.I. are becoming more common, even as questions linger about their psychological and ethical impacts. The scientists hope their findings might help ensure these technologies are developed responsibly.
“If we want to understand how to design these systems ethically, we need more evidence of this kind,” Tomasz Hollanek, a technology ethicist at the University of Cambridge who was not involved with the study, tells Science News.