Can an AI bot replace your therapist?

The healthcare industry has adopted artificial intelligence (AI) technology to diagnose disease, monitor patients, and even discover new drugs. 

But how can AI assist in the field of mental health, an endeavor flush with human emotion? Researchers are both hopeful and cautious on how AI can innovate mental health treatment.

Managing data with AI

In a review published in Current Psychiatry Reports, researchers wrote that mental healthcare has been slow to adopt AI technology but could benefit greatly from wider adoption, especially for parsing large data sets.^[]

Mental health practitioners generally rely on soft skills, such as forming relationships with patients and observing their emotions. AI has shown promise in combing through mounds of data to screen and diagnose mental illnesses.

“Leveraging AI techniques offers the ability to develop better pre-diagnosis screening tools and formulate risk models to determine an individual’s predisposition for, or risk of developing, mental illness,” the review authors wrote. “To implement personalized mental healthcare as a long-term goal, we need to harness computational approaches best suited to big data.”

Preventing suicide

In a paper published in the Australian & New Zealand Journal of Psychiatry, researchers used an AI predictive analytics tool to search old datasets.^[]

“There could be several advantages of incorporating artificial intelligence into suicide care, which includes a time- and resource-effective alternative to clinician-based strategies, adaptability to various settings and demographics, and suitability for use in remote locations with limited access to mental healthcare supports,” the authors wrote. 

An AI chatbot as your therapist

For example, there’s an AI chatbot service called Wysa, which prompts users with questions like “What’s bothering you?” then analyzes their responses.^[] The tool gives users supportive messages and advice—pre-written by psychologists—on managing issues like grief, depression, and chronic pain. 

In the 2021 Springer collection Multiple Perspectives on Artificial Intelligence in Healthcare, researchers wrote that chatbots have shown benefits in psycho-education and adherence.^[] But the authors also warn that chatbots could interfere with the patient-therapist relationship and may cause users to over rely on chatbots, which have demonstrated limited emotional intelligence. 

Defending the role of human therapists, authors of a 2021 paper published in SSM Mental Health argued that human interactions, not chatbots, must remain the first line of mental health treatment.^[] This is especially true for the most vulnerable populations, they wrote. 

Most Americans seem to bristle at the idea of using chatbots for their own mental healthcare.

Ethical guidance needed

Researchers warn that there must be guard rails to prevent misuse before AI is widely used in mental healthcare.

In an article about the ethical implications of AI in mental healthcare published in the Journal of Medical Internet Research, the authors suggested creating and implementing standard guidelines for AI in this field.^[] 

These authors also believed that the use of AI tools in mental healthcare should be transparent, with mental health professionals continuing to oversee the use of AI tools and algorithms. Clinicians also must avoid replacing traditional mental healthcare with robotic tools for those in need, they wrote. 

Can AI discern emotions in people’s facial expressions?

In addition to AI tools that dig into data and analyze human writing, there are now tools that purportedly recognize emotion in human faces. 

Ellie, a virtual avatar developed at the University of Southern California, uses nonverbal cues to guide the conversation, soothing users with human-like vocal intonations like “Hmmm.”^[]

Kate Crawford, PhD, a research professor at University of Southern California Annenberg and principal researcher at Microsoft Research, wrote in 2021 in Nature that the emotion recognition industry may be worth $37 billion by 2026.^[] Such projections run high, even though scientists disagree as to whether AI can detect emotions. In Dr. Crawford’s view, these tools need to be regulated to avoid being misused.

The future of AI in mental healthcare

While there’s hope that AI can be of great assistance to mental healthcare, it requires further research. Like with the AI devices that claim to detect emotions in faces, many of the AI tools in the mental health space remain under-studied. 

For example, a 2020 study found that only 2.08% of psychosocial and wellness mobile apps are backed by research.^[]

With more research, information, and ability will come a greater need for discernment. In an editorial published in the Journal of Mental Health, Sarah Carr, senior fellow in mental health policy at the University of Birmingham, wrote that the industry must work to consider the perspective of patients, caretakers, and families in how AI can best serve their mental healthcare needs.^[]

“If AI is to be increasingly used in the treatment and care of people with mental health problems then patients, service users and carers should participate as experts in its design, research and development,” Carr wrote.