Mental Health in the Age of AI

Renu Chaturvedi
Email: renu_chaturvedi@isb.edu
ISB Institute of Data Science, Hyderabad, India

World Health Organization estimated in 2020 that 20% of Indians suffer from some level of mental health issues. 56 million Indians suffer from depression and another 38 million are afflicted by anxiety disorders. India accounts for almost 37% of all suicides globally. Suicide is now the leading cause of death amongst teenage girls in India (aged 15–19). In addition to the social impact, mental health accounts for a huge economic loss to the tune of over $ 1 trillion from 2012–2030. (source: https://swachhindia.ndtv.com/world-mental-health-day-2020-in-numbers-the-burden-of-mental-disorders-in-india)

Now, more than ever, employees need availability and access to tools & means to restore mental wellbeing. In a pre-pandemic era, people would go out for a run, join a yoga class or reach out to a friend for help. In pandemic times, most people are homebound, have restricted travel, are working from home with added responsibility of care-taking and housekeeping responsibilities. This, coupled with the uncertainty the pandemic holds in present and in future in terms of availability of healthcare/ medicines/ hospital bed/ vaccines etc. has added to anxiety levels. News channels flooded with reports of human misery only serve to exacerbate the problem.

India has 0.3 psychiatrists and 0.07 psychologists per 100,000 population as opposed to the required number of 3.0. While we train more psychiatrists and psychologists to help tackle the problem, one potential solution lies in leveraging AI for mental health. A RAND study conducted in April 2021 found that the significant rise in tele-health use during the height of the pandemic was driven more by people looking for mental health services than physical health service (Erick Wicklund et 2021) With the advent of digital approaches to mental health, modern AI and machine learning is being used in the development of prediction, detection and treatment solutions for mental health care.

Chatbots like Wysa and Woebot offer conversational interfaces that facilitate well being by offering automated text interactions. While the technology is not made to replace a real therapist, it serves to encourage people to make appointments with a professional who can offer help and support. The other category of AI applications in mental health like Marigold Health, Mindstrong and Ginger offer ability to recognize behavior patterns in patients. For example, fitness devices track sleep patterns, pulse and heartrate amongst other parameters. Numbers beyond normal range could be indicator of a mental health issue. With consistent use, the device can help detect and monitor aberrations, if any. Early detection of a mental illness can prompt appropriate intervention.

All these AI and ML tools will not only benefit people who cannot afford therapy but also those who may be too shy or ashamed to try therapy. These will also benefit those living in remote areas with no access to a therapist. Thereby, reaching a broader base of population.

Despite improvements in awareness of mental health, there’s still stigma attached to it that prevents people from addressing it and seeking support. The anonymity that social media offers, allows more people to reach out for information, help and support. However, it is important that any mental health intervention on social media should be moderated by persons with the right credentials.

Social Media, almost ubiquitous in its use now can yield interesting insights in identifying early onset symptoms and perhaps even manifestation of a mental illness. An emerging field in research is that of ‘digital phenotyping’ that aims to study how individuals interact with their digital devices including their participation on social media platforms; in order to study patterns of illness. (Jain et al. 2015; Onnela and Rauch 2016). Machine Learning has enabled large volumes of data captured from social media platforms such as Twitter or Instagram to shed light on various features of mental health (Manikonda and De Choudhury 2017; Reece et al.2017). Conversations on Twitter have been analyzed to characterize the onset of depression. (De Choudhury et al.2013).

Nothing replaces the good old human connect, friendships, love, compassion and bonding over a meal or tea. In the absence of the physical connect in the pandemic times (and perhaps, post-pandemic too) and increased dependency on digital mediums for information, outreach and support; AI & ML will help generate insights into mental illnesses -its early detection and treatment & support.