AI-powered chatbots and apps are proving to be invaluable assets in the field of mental health, removing persistent hurdles to treatment such as cost, access, stigma, and therapist availability and fit.
A flash of panic, a wave of overwhelming despair: In the throes of a mental health incident, whether a chronic condition like depression or anxiety or a sudden moment of crisis, professional help can feel like a distant, difficult prospect. Who do you call? How will you find a therapist who is right for you? How long will you have to wait to see them? How much will it cost?
Rather than grapple alone with these questions, many people can now pull out their smartphone and engage immediately with a mental health app powered by generative artificial intelligence (AI). “How are you feeling?” the chatbot asks. “I’m here for you. Let’s figure it out together.”
As global demand for mental health services surges, AI-powered chatbots and apps are already proving to be invaluable assets in the provision of evidence-based psychotherapy such as cognitive behavioral therapy, breaking down traditional barriers to treatment such as cost, access, stigma, and therapist availability and fit. AI doesn't take days off, doesn't keep waitlists, never judges patients, and can offer help at a fraction of the cost of traditional therapy.
This article examines the current state of AI mental health tools, evaluates their potential for the assessment and treatment of mental health conditions, and explores some of the questions that remain around AI-powered therapy.
As an early and eager adopter of the internet as pathway for psychological treatment, I have spent much of my career designing, implementing, and evaluating digital mental health assessment and treatment programs with the goal of increasing access to evidence-based care. From my evaluations of some of the AI-powered mental health treatment apps in the rapidly evolving market, I’ve been surprised at how remarkably human they present as and how engaged I felt while chatting with their bots.
Even in these “very early days,” the benefits of AI mental health apps for patients are clear:
Human-like interactions: Despite being a machine, the sophistication of AI tools can convincingly replicate human conversation.
Accessibility and convenience: No waitlists or need to travel to a therapist’s physical office; digital therapy is always on and available.
Stigma-free: Many patients may feel less judged when interacting with a bot than a human therapist.
Evidence of AI's efficacy in mental health treatments is growing, with health systems like the U.K.'s National Health System (NHS) employing AI apps to handle an overwhelming number of referrals for mental health services. The NHS says these apps will help it respond to the 400,000 mental health referrals the system receives annually that are triaged but do not result in any face-to-face support. When patients interact with a chatbot during eTriage, the app flags high-risk responses for crisis attention from mental health clinicians and offers treatment and exercises to lower-risk patients.
Recognizing that mental health treatment gaps in Africa exceed global averages, the South Africa-based Vimbo app offers self-guided digital therapies for depression and anxiety. The app requires no human intervention, with its content, tools and technology acting as a non-drug cognitive treatment. This enables employers an opportunity to offer stigma-free support to their employees. Reported one Vimbo user, “I learned how to cope, learned how to relax and deal with issues as to not affect me the way they did before."
Woebot, which markets itself as a “relational agent for mental health,” is only available to users through health plans or systems, not as a standalone app for individuals. In addition to its original program for adults, it offers specialized AI chat tools for adolescents and mothers facing postpartum depression. Studies around Woebot have reported that users quickly establish a therapeutic alliance with its bot, and that the tool can effectively support users with substance use disorders, young adults with depression, and new mothers facing mental health challenges.
As AI continues to develop, the potential for these tools to support and transform mental healthcare is immense.
Quick access to treatment within a critical window of time can lead to meaningful treatment engagement. Literature suggests that most people have a six-week window of motivation to seek mental health services. If they can’t secure a first appointment within that window of time, motivation drops, and the chances of engaging a person in meaningful treatment are reduced. Connecting with a human practitioner is important, but so is responsiveness; the instantaneous support of an AI chatbot fills that need.
Successful outcomes in psychotherapy depend on matching people to the right evidence-based approach for their problem or condition as well as matching them to the right therapist based on “fit,” which can encompass culturally appropriate care and preferences for gender, cultural or ethnic background, LGBTQAI+ and virtual versus in-person care. The human-free nature of AI treatment tools could help overcome pervasive inequalities in access to mental healthcare.
AI tools could accelerate positive patient outcomes when combined with traditional therapy approaches. The traditional model of talk therapy, where a therapist meets with a patient weekly for one hour, is a social and economic convention that is likely not the optimal cadence for all people and problems. It is hardly “intensive.” Yet transformational change in conditions like chronic depression most often occurs with intensive treatment. Discussions and completion of homework exercises with an AI chatbot in between in-person sessions can elevate the intensity of the treatment and lead to better outcomes for patients. The psychology community would agree that we have enormous room to improve the treatment response for patients struggling with depression, obsessive-compulsive disorder (OCD), generalized anxiety disorders (GAD), and similar conditions.
AI can support our frontline helpers: the NHS has launched trials of its mental health app among healthcare workers, who reported improvements in mood, sleep, and overall wellbeing while using the program.
A patient struggling with an injury, cancer diagnosis, disability or substance use problem can benefit enormously from mental health check-ins and therapies with an AI chatbot.
Relapse is a chronic problem that generates huge costs for insurers. AI therapy chatbots could be a powerful tool in the fight against relapse by flagging people on the verge of relapse and reengaging them with treatment. For example, a chatbot could remind a patient of specific strategies that worked for them in the past, like talking to family members, limiting time on social media, and getting outside for movement.
AI tools could passively monitor patients by sending them a monthly questionnaire, gathering physical data via digital wearables, or tracking their social media accounts for concerning posts.
A combination of evidence-based psychotherapy and/or second-generation antidepressants remains the first line treatment for most common mental health conditions such as depression and anxiety, with research generally showing that psychotherapy alone is more effective for many conditions. Additionally, many medications for depression and anxiety produce unwanted side effects like weight gain and sexual dysfunction. Given the overall relatively modest effects of treatment and the challenge of understanding what works best for whom, AI could help tailor more effective combination treatments. For example, AI tools can evaluate patient preferences and medication side effects, and perhaps help some patients transition away from medicine alone to effective evidence-based psychotherapy.
AI chatbots can serve as a gateway to traditional therapy for individuals who may be hesitant about traditional in-person sessions. They can also engender positive associations for therapy among adolescents, who regularly deal in instant communication and feel more comfortable interacting with technology.
While human psychotherapists may approach AI mental health tools warily – fearful theirs is another profession susceptible to AI takeover – AI tools hold the potential to significantly assist practitioners and reduce administrative burdens. Therapists can employ AI tools to screen and triage patients, summarize session notes, and draft research and presentations.
AI tools can passively and actively capture patient data to create hypotheses that practitioners can test during in-person sessions. Psychotherapists could also use AI chatbots to train themselves or others on virtual patients.
AI therapy can also relieve the strain on human therapists, who themselves handle the mental and emotional load of serving many distressed people each day.
Despite its promise, the deployment of AI in mental health care isn't without its challenges.
AI offers a new horizon for mental health treatment, with significant potential for better access, enhanced care, and more robust outcomes. As technology evolves, the blending of AI with traditional methods may prove to be the fulcrum on which the future of mental health treatment balances, offering brighter prospects for patients and providers alike.
