Overview
- A Tel-Hai University survey of 584 Israeli generative-AI users found 29.3% used chatbots at least once for emotional support and that rate rose to 47% among those reporting severe psychological distress.
- Analysis of the Tel-Hai data showed the intensity of distress was the strongest predictor of turning to AI, with younger people and women more likely to seek help from chatbots.
- A separate OxWell survey of 24,132 adolescents in England found 2.7% used general-purpose AI for mental-health support and that 63.2% of those users said they wanted additional help.
- Both studies rely on self-report and nonrandom or regionally concentrated samples, and neither evaluates whether AI interactions were therapeutic, harmful, or substituted for professional care.
- Researchers and clinicians are urging routine questioning about patients' AI use, more research on safety and crisis response, and technical and policy guardrails because consumer LLMs lack clinical validation and mandated emergency protocols.