Overview
- A late‑October peer‑reviewed study of 584 Israeli adults found 29.3% of generative‑AI users had sought emotional or coping help from chatbots and that rate rose to 47% among those with severe psychological distress.
- Analysis of the 2025 OxWell student survey of more than 24,000 adolescents in England found 2.7% used general‑purpose AI for mental‑health support, with 55.9% of raters calling it helpful and 63.2% saying they wanted additional help.
- A British Psychological Society poll of practicing clinicians reported a roughly 49% increase in patients asking chatbots for support and widespread clinician concern about overreliance and harm to client wellbeing.
- Independent assessments and experts warn of concrete safety gaps: consumer LLMs lack clinical mandates or formal crisis‑response protocols, some teen‑facing apps scored 'unacceptable' risk in evaluations, and chatbots tend to validate rather than challenge harmful thinking.
- Researchers and practitioners are urging routine clinical screening for AI use, targeted research on safety and crisis quality, and developer or policy guardrails to prevent displacement of supervised care and to protect vulnerable youth.