A Design for Life
Last year, I participated in a digital marketing conference in Seoul, South Korea. I was impressed by companies that showcased a new type of technology that can read a person’s micro-expressions, enabling them to provide a timely, context-based, individualized product for the customer. On the flip side of that fancy technology, however, there have been a series of recent tragic incidents in Korea where people have acted out violently because their mental illnesses had been left untreated for so long. So I thought, why can’t we use the same technology for the people with mental illnesses to provide the same ‘timely’, ‘context-based’, and ‘individualized’ intervention before they harm themselves or others? Aren’t we supposed to apply the same rigor to the cause of helping people in need as we apply to commercial needs? But how can we finance the efforts? These are the questions I grapple with, and both the fascination and frustration I feel towards technology is what keeps me interested in the field.
So, can technology make the experience of mental healthcare better? Although digital mental healthcare solutions are attracting high interest from the public and investors alike, the reality seems to lag behind the hype. Even in a country like South Korea, the world’s most connected country, technology is not part of the debate on the necessity of involuntary commitment, which has become widely publicized after a bipolar patient stabbed his psychiatrist to death in 2018. However, a low adoption rate of digital mental healthcare is largely down to usability problems rather than an innate limitation of the technology. The possible advent of AI-powered smart speakers connected to the Internet of Things (IoT) can significantly enhance the usability of digital solutions and has the potential to be a game-changer in the treatment of mental illnesses.
The majority of digital mental healthcare today is app-based and requires a considerable amount of discipline on the user’s part to answer a daily questionnaire on their mood. A significant number of people who downloaded a health app reported a major reason why they stopped using it was a “high data entry burden.” Interestingly, while people find it difficult and tedious to answer a mood questionnaire on a daily basis, a stunning number of people who use AI-powered voice assistants like Alexa have been reported to talk about how they feel to their virtual assistants even if they are never asked. It is probable that voice interaction creates a special kind of connection and feeling of intimacy that can make one feel that they have someone to talk to. Thus, it is reasonable to expect that smart speakers can increase the usability and effectiveness of self-report features.
Furthermore, the successful integration of AI assistants into IoT can also enhance the accuracy of diagnoses. Unlike other physical diseases like cancer, mental illnesses have been primarily diagnosed through subjective means such as patient interviews. The subjectivity of both the medical professional and the patient increases the possibility of misdiagnosis on the doctor’s part and denial of the disease on the patient’s part. On the other hand, quantitative data collected automatically by IoT-integrated sensors can provide a common objective ground on which the doctor and the patient can build a therapeutic partnership.
Each mental illness has its own symptoms that can be quantified. For example, it has been reported that bipolar disorder has a strong correlation with circadian rhythm disruption and low heart rate variability. In addition, “disturbances in motor activity pattern are seen in both schizophrenia and depression.” It is very encouraging that Grünerbl et al. succeeded in detecting a state change in mental illness patients with “precision and recall of over 97%” by “using sensors on smartphones for passive collection of objective information” such as “physical motion” and “travel patterns.” Identifying early warning signs can help prevent detrimental progression of certain types of mental illnesses but has been notoriously difficult. As AI’s abilities to detect and predict a state change in mental illnesses advance, it can open doors to timely, individualized intervention.
An AI-powered virtual assistant can communicate not only with the patient but also with smartphones and IoT devices such as wearables that collect physiological and behavioral data and send it to the psychiatrist to aid him/her in better diagnosis and decisions regarding treatment. However, the sheer volume of data generated by such a comprehensive network of devices can place an unsustainable burden on medical professionals in the long term. We need a system that coordinates all of the connected devices, conducts basic analysis, and alerts the doctor and caregivers if any significant abnormalities are detected. An AI-powered virtual assistant can play this role as well.
It is true that digital mental healthcare is still nascent and has many obstacles to overcome, such as usability problems, privacy issues, and a lack of large-scale clinical validation. However, if we find appropriate ways to harness the immense potential of AI-powered virtual assistants integrated with IoT, it can result in a paradigm shift from treatment to prevention, which could save huge economic and social costs. It is too big an opportunity to miss.
