ShrinkGPT: Is AI Therapy the Answer? by Dr. Timothy Smith

Photo Source: Pexels

Recently published research on AI therapy’s efficacy suggests that chatbots make a difference for mental health patients. Chatbots such as Claude from Anthropic and CoPilot from Microsoft have proliferated widely since November 2022 when OpenAI released ChatGPT3.5. Chatbots have infiltrated automated help lines, smart speakers, medical practices, and many more businesses and institutions. Natural extensions of the chatbot include the mental health arena, where talk therapy plays an enduring role in clinical practice.

The need for broader access to mental health services continues to grow faster than the pool of licensed therapists. According to The Pew Center, a charitable trust, the United States has witnessed a growing number of its citizens with reported mental issues. According to Pew, “People in America have long grappled with mental health conditions and substance use, and for many individuals, the COVID-19 pandemic exacerbated these issues. The numbers are stark: In 2021, nearly 1 in 4 adults had a mental illness; nearly 1 in 3 adults had a substance use disorder or mental illness…” (pewtrusts.org). According to the American Psychological Association, as of 2018, the US had 31 licensed, doctoral-level psychologists per 100,000 Americans. (apa.org) These numbers translate to one therapist per 3,226 people in the US, far below the level needed to treat one in four people with mental health disorders today. Expanding the definition of mental health providers to include therapists raises the number of providers to 279,811 or 1 provider for every 1,191 Americans. (wellbeingport.com)

The significant shortfall of mental health providers today stems from many causes, such as the extensive training needed to become a licensed therapist, the high burnout rate for therapists limiting their practice to an average of fifteen years, and reimbursement challenges. The massive shortfall of mental health providers today suggests that alternative solutions need exploration, given the current situation. Various mental health chatbots have entered the market to address the shortfall of therapists with automation, and people are using general chatbots for mental health advice. The FDA has not approved any chatbots for mental health treatment. Still, products such as Therabot, Ada, Chai, and Mindspa have millions of customers using the apps for anxiety, stress management, depression, and sleep disorders, among others.

Just recently, a group of researchers at Dartmouth College undertook a randomized, national clinical trial of the effectiveness of a generative AI-powered chatbot called Therabot. Therabot treated patients with clinically significant symptoms of major depressive disorder, generalized anxiety disorder, or feeding and eating disorders. The trial consisted of 210 patients, half of whom received access to Therabot, and the other half were told they were on a waiting list for access to the chatbot. The study results were published on March 27, 2025, in the New England Journal of Medicine AI and titled “Randomized Trial of a Generative AI Chatbot for Mental Health Treatment.” The authors found significant improvement in symptoms for each of the three groups of patients compared to the control group. The patients also reported the therapeutic relationship with the chatbot as “comparable to that of a human therapist.” (ai.nejm.org)

These findings need additional research to add to their validity. Still, the implication that a chatbot tuned for mental health counseling demonstrated clinical efficacy will extend the reach of the overwhelmed community of mental health providers. The US suffers from a lack of mental health providers overall, but some states have far fewer mental health providers than others. States such as California, New York, and Massachusetts have many licensed psychologists. For example, Massachusetts has 81.9 therapists per 100,000 people, and states such as Mississippi have only 12.1 per 100,000 residents. The development of better chatbots for mental health accessible by phone or internet could significantly broaden the reach of mental health therapy not possible with the current system.

Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high-level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.

You can buy his book on Amazon in paperback and in kindle format here.