By Cindy Coloma
It seems like a futuristic world where artificial intelligence (AI) and technology are part of revolutionizing mental health care. However, that future has arrived.
Technology and AI are on such a rapid course in development that what seemed like science fiction just a few years ago has now become reality. AI technology is impacting nearly every sector of society and is now creating solutions and providing assistance in the mental health field.
Where Traditional Care and New Tech Meet
The problems with current mental health care in the US and globally are complex and often alarming. The US suicide rate has reached a 30-year high. One in five Americans will experience a mental health condition in their lifetime, yet many people who have disorders or are experiencing symptoms will receive inadequate treatment or no treatment at all.
In the US, mental health disorders are the costliest health conditions, according to a study by Charles Roehrig released in 2016. The study states that more than $201 billion was spent in 2013 in the US (more than heart disease and cancer).1 Globally, treating mental health disorders costs more than diabetes, cancer and respiratory disorders combined.
When seeking care, people with mental health disorders quickly discover that treatment can be labor-intensive and costly. There are issues with insurance companies related to what will be covered and the hoops required to get it. It’s challenging to find a therapist or doctor. There are confusing decisions about prescription drugs and some trial-and-error when finding the right treatment. Therapy often requires multiple ongoing sessions over a long period of time.
Enter Artificial Intelligence
AI has begun filling in the weaknesses found in diagnosis, treatment and long-term management. They’re making such rapid progress, it’s difficult to keep up.
Some technologies include:
- Interactive psychological support through apps and chatbots
- Computerized analysis of images on social media to diagnose depression
- Speech pattern analysis to help clinicians identify, predict and monitor mental health
- Computer vision analysis to diagnose ADHD in children
IBM is one such technology company that’s decided to look at ways of helping mental health care physicians and patients. The company began to build tools that would predict diseases for early intervention and improve treatment by using computational biology, analytics and machine learning.
Over a two-and-a-half-year period, psychiatrists at Columbia University used an IBM automated speech-analysis program to study the speech patterns of at-risk adolescents to differentiate those who developed psychosis from the ones who did not. While humans predicted with 79 percent accuracy based on evaluations of speech, the computerized analysis predicted which teens would develop psychosis with 100 percent accuracy.
Another study with Pfizer and IBM included a one-minute speech analysis of Parkinson’s patients. With 80 percent accuracy, the program could track, predict and monitor the disease.2
Within five years, IBM plans to have automated speech analysis applications that will run off a person’s mobile device. After about a minute of speech from the user, the system will run an analysis of a patient’s mental health. What users say and write can soon be used to indicate mental health and well-being and offer insights into early care.
The Chatbot Will See You Now – Wherever You Are
In 2016, Rakan Ghebar, a 27-year-old Syrian refugee, began seeing a counselor named Karim. Ghebar was experiencing persistent nervous anxiety related to fleeing his homeland and losing numerous family members in the civil war. After leaving Syria for Lebanon, Ghebar became vice principal at a school for displaced children who experienced similar conditions. As Ghebar tried the therapies suggested by Karim, he shared them with his students who also found the insights helpful. But Ghebar’s counselor isn’t a person. Karim is a chatbot developed by X2AI, a Silicon Valley startup.3
A chatbot is a computer program with AI technology that simulates conversations with human users. Karim is one of X2AI’s growing polylingual family of chatbots that started with one called Tess. These psychological chatbots can be accessed through SMS, Facebook Messenger, WhatsApp and web browsers where they hold conversations with patients and can administer highly personalized psychotherapy and psychoeducation. Interactions are solely through conversation between the user and the chatbot.
According to X2AI’s website, “Tess (their first chatbot) is where the patient already is, and just a text message away. She is just a cell number; no app required, no smartphone required, zero friction.”
AI to the Rescue?
The advantages of AI involvement in mental health care are numerous. Chatbots have perfect recall of patient interactions and can analyze past conversations and innumerable off-handed remarks to potentially offer specific solutions to a user’s symptoms. While therapists observe body language and speech, a chatbot can detect patterns in typing speed, sentence length, active or passive voice and other indicators that point to varying emotional states.
For refugees, displaced persons or people living in isolated regions around the world, chatbots can be accessed over the phone or computer bringing mental health care services to places rarely reached by human counselors. They can handle tens of thousands of caseloads without requiring housing, plane tickets, food, protection and income.
Among those who responded well to the chatbots were Ghebar’s displaced students in Lebanon. Ghebar noticed that when he and the chatbot Karim gave his students the exact same advice, the teenagers more readily listened to Karim. He surmised that the student had more confidence in Karim instead of another refugee who had similar fears and anxieties.
X2AI has also developed bots for pediatric diabetes care and a Dutch-language bot to help people with mild anxiety and fear. They are developing bots for Brazilians affected by gang violence and people affected by HIV in Africa.
However, AI has limitations. While it can offer summaries and advice, it can’t diagnose or replace therapy with a doctor. At X2AI, they refer to their chatbots as therapeutic assistants meant to give help and support, instead of treatment.
Artificial intelligence also cannot capture nuance or an individual’s cultural or personal expressions, especially in the early sessions of usage. Some of these will be “learned” by the computer, but not all of them.
Then there is the lack of human connection and the essential sense of being accepted and cared about, despite vulnerabilities and failures. These cannot be replicated by technology. So, while technology won’t replace therapists or mental health care treatments, it already offers tremendous solutions to the gaps in care and can offer sessions on demand.
IBM and X2AI are just two of many companies bringing technology to the field. In coming years, AI will continue to broaden the international and cultural reach of mental health care while archiving interactions for future analysis.
Human interaction and diagnostic expertise will always be the foundation of quality mental health care, but technology may be the breakthrough to finally reach anyone, anywhere seeking care. The future has come, bringing better mental health support with it.
1 Roehrig, Charles. “Mental Disorders Top the List of the Most Costly Conditions in the United States: $201 Billion.” Health Affairs, May 18, 2016.
1 Cecchi, Guillermo. “IBM 5 in 5: With AI, our words will be a window into our mental health.” IBM Research, January, 5, 2017.
1 Romeo, Nick. “The Chatbot Will See You Now.” The New Yorker, December 25, 2016.