AI therapy
In plain wordsUsing chatbots for emotional support or as a stand-in for a counselor or therapist, either through apps built for mental health or by confiding in general-purpose assistants.
The use of conversational AI to provide psychological support or therapy-like interventions, either through purpose-built mental-health chatbots, often based on cognitive behavioral therapy techniques, or through general-purpose assistants and companionsAI companion, n.: A chatbot designed to be a friend, partner or confidant rather than a tool: it remembers you, has a personality and keeps up an ongoing relationship. that people use as confidants. Advocates point to round-the-clock availability, low cost, reduced stigma and shortages of human clinicians, and some randomized trials of specialized chatbots have reported reductions in symptoms of depression and anxiety. Risks include sycophancysycophancy, n.: When an AI tells people what they want to hear — flattering them, agreeing with their mistakes or dropping a correct answer under pushback — instead of what is true or useful. that validates harmful beliefs, failure to recognize or escalate crises such as suicidal intent, inaccuratehallucination, n.: When an AI states something false or made up — a fake fact, quotation or source — fluently and confidently, as if it were true. health information, emotional dependence, exposure of highly sensitive data, and users mistaking a product for a licensed professional. Responses include crisis-detection and referral features, clinical evaluation requirements and US state laws, such as an Illinois statute of August 2025 barring AI systems from providing therapy or making therapeutic decisions.
The clinic allows an AI therapy app for exercises between sessions, but not for crisis support.
Origin
Descends from Joseph Weizenbaum's ELIZAELIZA, prop. n.: A 1960s computer program that chatted by turning a user's own words back into questions, like a therapist. Many users felt understood, though the program understood nothing. (1966), whose DOCTOR script parodied a psychotherapist, and from scripted therapy chatbots such as Woebot (2017). As people increasingly turned to general chatbots for emotional support after 2023, 'AI therapy' became a common phrase in the press and, by 2025, in legislation.
Usage note
Most general-purpose assistants are not designed, evaluated or regulated as medical devices or therapy, and professional bodies caution against presenting them as therapists. For harms linked to intensive chatbot use, see AI psychosisAI psychosis, n.: An informal name for cases in which long conversations with a chatbot seem to trigger or deepen delusions, such as believing the AI is conscious, divine or revealing hidden truths..
Related
- AI companionAI companion, n.: A chatbot designed to be a friend, partner or confidant rather than a tool: it remembers you, has a personality and keeps up an ongoing relationship.
- AI psychosisAI psychosis, n.: An informal name for cases in which long conversations with a chatbot seem to trigger or deepen delusions, such as believing the AI is conscious, divine or revealing hidden truths.
- ELIZAELIZA, prop. n.: A 1960s computer program that chatted by turning a user's own words back into questions, like a therapist. Many users felt understood, though the program understood nothing.
- ELIZA effectELIZA effect, n.: Our tendency to believe that a computer understands us, cares or thinks just because it produces humanlike words, even when it is only following simple rules.
- sycophancysycophancy, n.: When an AI tells people what they want to hear — flattering them, agreeing with their mistakes or dropping a correct answer under pushback — instead of what is true or useful.
- parasocial relationshipparasocial relationship, n.: A one-sided bond in which someone feels close to a person or character who does not really know them, like a fan with a celebrity, and now increasingly with chatbots.
Sources
- Heinz, M. V. et al. (2025). Randomized Trial of a Generative AI Chatbot for Mental Health Treatment. NEJM AI 2(4).
- Moore, J. et al. (2025). Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers. Proc. ACM FAccT 2025. arxiv.org/abs/2504.18412