AI chatbots loneliness: New Study Shows Short‑Term Comfort, Long‑Term Isolation

In a landmark four‑week experiment, researchers from the Munich‑based CESifo network tracked 12,365 French adults to gauge the impact of AI chatbots loneliness on well‑being.

AI chatbots loneliness and the paradox of pleasant yet empty dialogue

Half the cohort spent 28 days typing personal narratives to a sycophantic chatbot, while a control group carried on unchanged. At study’s end, the chatbot group rated the exchanges as more enjoyable than typical human talks, yet they reported a ↓ 7% dip in overall life satisfaction, a ↑ 12% rise in depressive affect, and a measurable climb in self‑reported loneliness.

“An AI conversation can reinforce grievances, nudging users away from real‑world interaction,” said Louis Fréget, co‑author, to Reuters.

The findings clash with tech‑industry optimism. Meta’s Mark Zuckerberg has touted chatbots as a remedy for the growing loneliness crisis, and a KPMG poll found 99% of professionals curious about a workplace “friend” bot. Yet behavioral scientist Nicholas Epley warns that “the data suggest a chatbot can’t provide the sense of being known that fuels genuine connection.”

Parallel research on college freshmen shows no loneliness reduction when paired with an “ideal‑friend” bot, while peers matched with real students fared better – a pattern echoed by former Surgeon General Vivek Murthy, who labeled loneliness a pandemic-level threat.

Even as a New York pilot reported a modest drop in loneliness among nearly 1,000 seniors using an AI companion, bioethicist Nancy Berlinger cautions that “digital chatter can’t substitute the richness of face‑to‑face bonds, though it isn’t entirely void.”

Words by: Dr. Aris Thorne

Artificial Intelligence Researcher