The pandemic has greatly increased the demand for online health treatment, especially with regard to mental health issues. Studies suggest that 1 in 5 Americans will face a mental health condition within their lifetimes – and the unprecedented socioeconomic changes made in response to the pandemic could result in similar outcomes worldwide. A study from the Department of Mental Health in Thailand in 2012, for instance, showed that approximately 20% of the population struggled with some form of mental illness. This situation didn’t get any better recently, as Thailand recorded 7.37 suicide deaths per 100,000 people in 2020 alone.
Post-COVID challenges in the mental health space include deteriorating economic opportunities at a time of great social upheaval, the attrition of medical staff as a result of overwork and burnout, and the enduring stigma surrounding mental health support. In such an environment, the need for innovative solutions is urgent.
Demographic shifts have also led people to move away from city centers, and many are now struggling to access the resources they need. Thailand’s Mental Health Department currently has 20 telephone counseling hotlines in operation – hardly enough for a country of 69 million people.
Fortunately, there is great potential for AI within the mental health arena. As we will see, the proper integration of AI technology enables more effective and personalized treatment plans – both for therapists and their patients. Here we outline three ways that AI can be used to improve mental health therapy.
AI-powered apps for mental health
One of the fastest growing areas of concern regarding mental health relates to depression and anxiety disorders. New research on emergent AI technology has shown great promise in this area, with a recent study finding that AI monitoring through motion sensors and deep learning is able to detect anxiety symptoms with over 92% accuracy. For example, Chulalongkorn Hospital, in collaboration with Thailand’s Ministry of Public Health has launched DMind, an app that uses deep learning to analyze facial expressions, voices and messages in order to diagnose people with depression.
Moreover, AI chatbots that can carry a daily conversation with users, ask basic questions, and present information in a consistent tone have been shown to create a positive therapeutic experience.
Though it may sound surprising, people seem more comfortable talking to a “robot” than a therapist about mental health issues. A recent survey conducted by Oracle and Workplace Intelligence found that just 18% of respondents would rather talk to a human about their mental health struggles, while 68% would prefer a robot. AI and technology is commonly seen as less biased, less judgmental, and faster to provide feedback than traditional human therapists. Perhaps ironically, people simply find it easier to open up to a human-like AI rather than an actual human.
‘Emotionally intelligent’ AI chatbots like Wysa and Youper use AI to react in real time to the emotions that patients express, which may often be more comforting than talking to an actual therapist. These and other chatbots provide privacy and anonymity, which can be extremely helpful for those who are struggling with the stigma of going to an actual therapist.
Natural language processing (NLP)
As part of the larger AI landscape, natural language processing (NLP) allows computers to analyze, understand, and derive meaning from both text and speech in a way that is similar to how humans handle such information. While every person’s mental health challenges are unique, a growing body of research indicates that statistics, speech analysis, and predictive behavior models are very effective ways of assessing potential as well as developing mental health issues.
One proof of concept study was able to predict psychosis 100% of the time when analyzing baseline and follow-up interviews. The transcript data from these interviews was fed into an algorithm, which in turn reviewed speech coherence and complexity for evaluation.
AI can also use NLP to automate quality control among therapists, ensuring high standards of care for each patient. As staff turnover, burnout, and limited budgets squeeze existing mental health professionals, the likelihood of people “slipping through the cracks” or warning signs being overlooked becomes greater without such measures.
NLP also provides an effective way to quickly and efficiently screen patients for specific types of care. As a low-cost treatment option available 24 hours a day, AI powered by NLP can be a valuable tool that either replaces or augments traditional human therapists.
The move toward telehealth and virtual therapy has generated massive amounts of raw metadata on treatment options and outcomes. Machine learning uses computer algorithms to detect patterns inside this huge trove of information. By learning from historical data, its analytical tools can automatically spot mental health issues and predict future outcomes among new patients. This technique can be deployed widely, and can make up for any shortfall in the number of mental health care professionals currently available.
Machine learning can also figure out which type of therapy would work best for prospective clients, and help match them with the right therapists. With machines working alongside trained professionals, the human therapists can focus on special cases while the remaining majority of mental health discussions can be performed with the help of AI. Such a process would increase overall efficiency, speeding up recovery for most patients and providing more tailored care when needed.
The way forward
The use of technology in mental health treatment is at a revolutionary crossroads. AI-based apps and programs are able to amass data, diagnose conditions, track moods, and provide care. They make assessments using games, text-analysis software, and predictive modeling – able to serve an almost unlimited number of clients both instantly and simultaneously.
With the help of AI, mental health therapy can be always at the fingertips of those who need it, while also improving in terms of affordability, accessibility, efficiency, and privacy. As these AI-enhanced technologies are put to further use, its algorithms will have greater access to anonymized user data, and will therefore be better equipped to guide patients to more favorable outcomes.
Despite the challenges that the pandemic has created over the past two years, there is ample reason to be hopeful. As the world’s people and economies recover from COVID, AI-powered therapy sessions can now enhance people’s mental health journeys, leaving them better prepared to move forward with a healthy frame of mind.