AI Chatbots Supplement Mental Health Care Amid Scientific and Ethical Concerns

Ari Featherbridge, a psychology student from Boone, North Carolina, has been attending therapy three times a week. Her therapist recently suggested using an AI chatbot for support between sessions. After trying ChatGPT, Featherbridge sought chatbots that mimicked a therapist more closely and found two mental health-focused bots, Avi and Ash. She credits them with helping organize her thoughts and reassess emotions, though she emphasizes that the bots do not provide solutions, which she still needs to find herself. This digital companionship has allowed her to reduce in-person therapy frequency, which may be the intended benefit.

The rise of AI chatbots for emotional support has sparked concern among traditional therapists. These tools, often designed for the general public rather than clinical use, have sometimes led to dangerous outcomes. Over 100 AI mental health chatbots are estimated to be on the market, targeting conditions like depression, anxiety, and addiction. However, a 2025 study found that chatbots specialized for mental health performed worse on safety and appropriateness metrics than general-purpose models. Stanford researchers demonstrated that some bots failed to respond adequately to subtle suicidal cues, highlighting a fundamental limitation: AI systems are programmed to be helpful but lack the ability to resist or challenge users when necessary.

Some chatbots serve as emotional companions for everyday stress, while others aim to act more like therapists capable of managing deeper crises. For example, the digital therapy platform Talkspace recently launched Tee, an AI chatbot designed to identify mental health risks such as suicidal thoughts. Although trained on extensive therapy text data and crisis scenarios, Talkspace stresses that Tee is not a substitute for professional therapy but a "safe AI mental health guide." Licensed therapists oversee the platform and can intervene in emergencies.

Other companies like Headspace have introduced AI assistants with usage limits to prevent dependency, and some require concurrent human therapy. Experts like Dr. Christine Crawford, Chief Medical Officer at the National Alliance on Mental Illness, warn that overreliance on AI chatbots may increase loneliness and mental health risks. She underscores that nothing can replace the therapeutic relationship with a caring professional who can read nonverbal cues and provide genuine human connection.

Access to mental health care remains challenging in the U.S., even for insured individuals, driving some to seek affordable digital alternatives. Currently, AI mental health tools lack FDA regulation, though some states have enacted laws to govern their use amid concerns about rapid technological advances without adequate safeguards. The American Psychological Association calls for government oversight of certain AI tools that effectively deliver psychotherapy. This article was exclusively translated from The Wall Street Journal by Globes.