ai
Student Texts With Peers Cut Loneliness While Chatbot and Journal Groups Show No Change
@NewScientist reported that a University of British Columbia study found daily chatbot conversations produced no change in loneliness scores, whereas exchanges with fellow students lowered them.
New ScientistAudio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
A two-week trial led by Ruo-Ning Li at the University of British Columbia found that students who exchanged daily messages with a chatbot named Sam reported no overall change in loneliness, while those who texted randomly assigned fellow students felt less isolated.
Researchers built Sam on ChatGPT-4o mini and instructed it to act as an empathetic roommate. ” Around 300 students were split into three groups.
One third sent at least one message each day to Sam, another third exchanged daily texts with a peer, and the final third kept a one-sentence journal. Participants averaged eight to 10 messages per day across the trial. ” The group that texted other students recorded lower loneliness scores after two weeks.
The chatbot group and the journal group showed no measurable shift, leaving participants in both conditions feeling equally disconnected. ” The trial builds on earlier observations of human attachment to conversational programs. In the 1960s, Joseph Weizenbaum created the ELIZA algorithm, which turned each user statement into a follow-up question.
Weizenbaum’s secretary found the sessions so intimate that she asked him to leave the room during her use. U.S. adults and 25 per cent of people under 30 have used AI for companionship.
