Is AI therapy effective?
It depends what you mean by effective. As a substitute for clinical care, no. As a companion that helps you reflect, see patterns, and stay connected to your own inner life — yes, and the early research is encouraging. The honest answer lives between the marketing and the dismissals.
By The Calm AI Therapy editorial team. Not reviewed by a licensed clinician, and not a substitute for one. How we research and write.
What the evidence says
Studies on conversational AI tools for mental wellness show measurable reductions in self-reported stress, anxiety symptoms, and rumination — particularly when the tools are used between sessions with a human clinician, or when human care isn't accessible. The effect size isn't dramatic, but it's real and replicable.
Where AI therapy actually helps
Continuity. The biggest reported benefit isn't the response in any single session — it's having a place to think, anytime, that remembers you. People with anxiety report fewer 3am spirals. People between human sessions report bringing more clarity to those sessions. People who can't access human care at all report finally having something.
Where it doesn't help
Severe mental illness, acute psychosis, complex trauma, anything requiring medication — these need a human clinician. AI is a companion, not a treatment. A good AI therapist will tell you this directly when it sees something that requires professional care.
Common questions
Can AI therapy replace a human therapist?
No. It can be a companion alongside human care, or a useful first step toward it. For serious or persistent issues, see a clinician.
How long until I notice an effect?
Many people notice more clarity after the first few conversations. The compounding benefit comes from continuity over weeks: Aura remembers, so each conversation starts further along.
Is the evidence for AI therapy peer-reviewed?
Some early trials are. The field is young; expect more in the next two to three years.
Sources
- Baumel et al., Objective User Engagement With Mental Health Apps — Journal of Medical Internet Research, 2019
- The Supportive Accountability Model in Digital Health: a scoping review — Journal of Medical Internet Research, 2025
- "This app said I had severe depression": unintentional harms of mental health apps — ACM CHI, 2024
- APA advisory on the use of AI in mental health — American Psychological Association, 2025
Open a space. See whether it feels useful for you, specifically.
Try one sentenceRelated