by Jennifer Chiou
11th grade (Parkland, FL)
Special Recognition
During a time characterized by what seems like a never-ending stream of disasters, people across the globe are struggling to stay afloat––and not just physically. Currently, mental health problems are the single largest cause of disabilities in the world, acting as a major barrier to global health and wellbeing (PAHO 2019). Despite this, the most vulnerable people are often left without resources to alleviate their struggles. Everywhere, lower income people face obstacles to psychological care––children are given criticism instead of care for their ADHD, teens are forced to grit their teeth through their anxiety, and adults are left to drown in the effects of untreated trauma. In fact, low-income countries have an average of 0.1 psychiatrists per 100,000 people (WHO 2016), making it almost impossible for citizens to receive proper treatment.
Amidst this sea of unrelenting turmoil, software engineers are turning the tides with an innovative solution to mitigate the shortage of mental health care by developing artificial intelligence systems that virtually deliver cognitive behavioral therapy (CBT), a therapeutic method centered around reframing cognitive distortions. These systems utilize a subset of machine learning called natural language processing to essentially create a virtual therapist that can perform complex sentiment analysis and dialogue-belief tracking, mimicking a human psychologist.
One popular implementation of AI therapy is the “Woebot” chatbot created by psychologist Alison Darcy. This model was trained with thousands of examples of user input in order to differentiate between emotions, thinking errors, and logic, thus recognizing and responding to a user’s emotions almost exactly how a real therapist would in a texting/chatting format. But the unique aspect of Woebot is that it’s completely automated, requiring no human intervention whatsoever. A team of engineers debated on this feature, as human augmentation could improve Woebot’s performance, but they decided to prioritize user anonymity by making Woebot respond based on data given by each individual user.
In addition to increasing privacy and combating internalized social stigma against seeking psychological help, this also means anyone with a mobile device will have 24/7 access to mental health support because it eliminates the difficulty of finding a physical mental health clinic and ensures support at all hours of every day. This is imperative because often, even when someone can afford health care, they might not receive it due to schedule conflicts and overcrowdedness. For instance, one high-achieving 11th grader named Grace was hospitalized while experiencing a panic attack that caused her to self harm. The hospital searched her entire state, but all of the facilities that could treat her were full, and after four days, she had to go home without treatment (Star Tribune 2021). Unfortunately, Grace’s story isn’t an exception––enduring immense suffering without anyone being able to help is the reality for a vast majority of people with mental illness.
But AI has the power to change that. The efficacy of Woebot’s services has been proven in a large-scale study with over 36,000 users aged 18-72, which found that Woebot creates lasting therapeutic bonds that are non-inferior to the bond created between human therapists and patients (JMIR Formative Research 2021). AI and web-based CBT in general has also been shown to reduce levels of depression and anxiety, and although it isn’t perfect, it has increased accessibility to therapy, especially for lower-income citizens who are more likely to have access to a cell phone or computer than to mental health care.
However, it is important to consider that a digital divide exists in the world and people in very rural areas do not have access to electricity or network connectivity, so these structural and social barriers must be overcome to reach the most isolated individuals. Artificial intelligence is not the only solution to the global mental health care deficiency, but it is certainly a stride towards the greater accessibility and revolutionary quality of care that will have an extraordinary ripple effect on the welfare of generations to come.
Bibliography
Brooks, J. (2021, March). Grace note: A teen shares the story of her long wait for help. The StarTribune. Retrieved February 1, 2022, from https://www.startribune.com/grace-note-a-teen-shares-the-story-of-her-long-wait-for-help/600035581/
This news article shared the story of Grace, a junior in highschool who struggled with an intense panic attack and self harm, but was not able to receive any treatment due to space shortages in her home state of Minnesota. It was extremely touching and sad, emphasizing the deficiencies present in the mental health care industry and the social stigma surrounding psychological treatment.
Darcy, A., Daniels, J., & Wicks, P. (2021, May). Evidence of Human-Level Bonds Established With a Digital Conversational Agent: Cross-sectional, Retrospective Observational Study. JMIR Publications. https://doi.org/10.2196/27868
This report detailed the effectiveness of the Woebot chatbot's implementation of Cognitive Behavioral Therapy in comparison with that of a human therapist, finding that Woebot was able to create similar therapeutic bonds with users as a real therapist would. This conclusion was made based on 36,070 users, aged 18-78.
Fitzpatrick, K. K., Darcy, A., & Vierhile, M. (n.d.). Delivering Cognitive Behavior Therapy to Young Adults With Symptoms of Depression and Anxiety Using a Fully Automated Conversational Agent (Woebot): A Randomized Controlled Trial. JMIR Publications. https://doi.org/10.2196/mental.7785
This study tested the effectiveness of the Woebot chatbox in delivering Cognitive Behavioral Therapy (CBT) to college students who believe they symptoms of anxiety and development. It maintained that typical web-based methods of delivering CBT have been effective, but less accessible and harder for users to stick to, and found that the participants of the study saw significant reductions in their depression levels when using the fully automated Woebot.
Fulmer, R., Joerin, A., Gentile, B., Lakerink, L., & Rauws, M. (2018, October). Using Psychological Artificial Intelligence (Tess) to Relieve Symptoms of Depression and Anxiety: Randomized Controlled Trial. JMIR Publications. https://doi.org/10.2196/mental.9782
This study gave college-aged participants access to Tess, a psychological AI trained in Cognitive Behavioral Therapy, and found that usage of Tess led to statistically significant reductions in anxiety levels.
Mental Health Service Availability. (2016). https://www.who.int/data/gho/data/indicators/indicator-details/GHO/psychiatrists-working-in-mental-health-sector-(per-100-000)
This page contained a wealth of data about mental health availability, including the numbers of psychiatrists, mental hospitals, and psychologists per 100,000 people in each country.
Naslund, J. A., Aschbrenner, K. A., Araya, R., Marsch, L. A., Unützer, J., Patel, V., & Bartels, S. J. (2017). Digital technology for treating and preventing mental disorders in low-income and middle-income countries: a narrative review of the literature. The lancet. Psychiatry, 4(6), 486–500. https://doi.org/10.1016/S2215-0366(17)30096-2
This source explained the feasibility and efficacy with which digital mental health care could help prevent mental disorders in low and middle income countries. It showed that although digital mental health care does improve access and quality, there are still barriers that bar people from accessing technology in the first place in very isolated regions.