Scientists observe striking link between AI chatbots and psychological distress (www.psypost.org)

A large cross‑national survey published in the Journal of Social and Personal Relationships found that people who use social chatbots (examples named included Replika and My AI) are consistently younger and report higher psychological distress than non‑users. Researchers analyzed autumn 2023 data from the longitudinal “Self & Technology” study covering six European countries (Finland n=1,095; France 1,014; Germany 900; Ireland 588; Italy 1,099; Poland 967). Chatbot use (binary), frequency of face‑to‑face contact, attitudes to technology, and demographics were collected alongside validated measures: the Mental Health Inventory for distress, a short UCLA loneliness scale, and a single‑item self‑esteem measure. Across all countries chatbot users showed greater symptoms of anxiety/depression; loneliness was higher in France, Germany, Italy and Poland but not Finland or Ireland, while self‑esteem effects were mostly neutral (one positive association in France). Use prevalence ranged ~8.7%–17.9%. Technically robust controls bolstered the associations, but the cross‑sectional design prevents causal claims—distress may drive chatbot uptake rather than chatbots causing poorer well‑being. The pattern suggests social chatbots often function as “weak ties” that provide conversation or distraction but may not substitute for deep social support. For AI/ML practitioners and researchers this highlights the need for longitudinal evaluations, richer behavioral and motivation data, and design focus on therapeutic efficacy, safety, and ethical deployment if chatbots are to be positioned as mental‑health tools.