An AI Bot Holds Up a Mirror to Mental Health Crisis During COVID
There has been a 52% increment in online searches related to mental health since the onset of Covid in India.
Right after the first national COVID lockdown VK, a longtime client of mine sent me a meme that showed the Mona Lisa with a gas mask and a revised title—“Corona Lisa”. VK has struggled with acute OCD for nearly all their life and memes are their way of dealing with bouts of anxiety. Rigorous cleaning practices, compulsive ritualisations around washing of hands, and accompanied mysophobia (a generalised fear of germs linked to dirt, unclean places, and contamination) are all a part of their fatiguing manual for self-preservation.
Often these behaviours and underlined beliefs that guide them have elicited ridicule and unsympathetic criticism even from those closest to them. In a recent conversation, they pointed out how in a world struck by COVID-19, observances similar to their own have been normalised and accepted with fewer critiques because they obviously help fight an invisible enemy and this felt dissonant to them.
Clearly, COVID has left a detectable spoor across all our psychosocial terrains.
The understanding of VK’s dilemma is not a precise division of acceptable and unacceptable responses to threats. The nature of a threat and its relational objectivity matter as well. OCD is often characterised by difficult-to-pinpoint or even irrational fears and worries whereas a raging pandemic that is claiming lives on a daily basis is anything but amorphous or disembodied.
The psychological and somatic coalescing during a major public health crisis isn’t new. Karl Menninger, an American psychiatrist, published a paper in 1926 outlining cases of people who had experienced influenza and were subsequently admitted to the Boston Psychopathic Hospital with acute psychoses. Of 80 people with full data available, 35 had an illness fulfilling criteria for a narrow definition of schizophrenia, “dementia praecox”, as opposed to delirium and other psychoses.
This was at a time when mental illness was still a massive slur and patients were treated rather inhumanely. Yet, today, as we are collectively trying to outlast a pandemic again, similar mountains need to be scaled. Sciensano, the national public health institute of Belgium, recently released its 5th health survey during the pandemic based on contributions from nearly 30,000 members of the public in Europe. Their data indicates that 40% of people now (Nov-Dec) feel that they receive inadequate social support, compared to 30% in September. Additionally, 64% of people aged 18 and over are unsatisfied with their social contacts, compared to 34.5% then. (Cited)
These are alarming increases.
Half of the Population Facing Psychological Impact of the Pandemic
In a paper published in Indian Psychiatry, the following results give us a pause:
A recently concluded systematic review and meta-analysis on the prevalence of psychological morbidities among the general population, health-care workers, and COVID-19 patients amidst the COVID-19 pandemic reported that about half of the population faced psychological impacts of the COVID-19 pandemic. Poor sleep quality (40%), stress (34%), and psychological distress (34%) were the most commonly reported problems across various studies. An online Indian survey has reported that about 40.5% of the participants reported anxiety or depressive symptoms. About three-fourths (74.1%) of the participants reported a moderate level of stress, and 71.7% reported poor well-being.
When we assess components of caste, gender, race, and sexuality in mental health studies, we often find that social stressors are exacerbating underlined mental illnesses.
There is little to no work is done across the board in India to address how folks from oppressed castes, particularly those working in gig economy setups or as migrant labourers, are dealing with their acute trauma at the moment. The help available is sparse and sporadic at best.
COVID 19 has altered our understanding of physical and psychological safety in both linear and discursive ways. It has pushed us to reconcile with social and economic inequities that can no longer remain concealed underneath ignorance like mold growing beneath an expensive house’s rafters. The direct impact of prolonged social isolation, the threat of job-loss, anticipatory grief around the health and wellness of loved ones who run the risk of contracting the illness are just some of the more evident ways in which people are still reeling from this pandemic.
There is destabilising precarity surrounding its eventual decline, ongoing disruptions to any sense of sanity, and the daunting task of rubbing eyeballs with our own mortality in its more perceptible form as opposed to some terminal, distant coin flip. Clients in therapy have been displaying heightened anxiety about survival– physical, financial, emotional, and social. And these are the people who can actually afford to get help without rationing resources for food and water.
Indians Were Searching For Mental Health Help
Our collective fragility is more transparent now than it was before. Most of this year, our lives have been lived online more than ever before. This is why I was intrigued when Angad Chowdhry—anthropologist, researcher, and co-founder at Quilt.AI tweeted about his organisation’s informative, AI-aided research about online trends related to mental health in wake of COVID-19. This research trained its lens on Indian demographics, particularly the urban populace.
The Quilt.AI team extracted data from 200 search terms and 100 unique posts from public fora as part of their study. Their guiding principle was to test what people are truly experiencing beyond the mask of social desirability.
They accomplished this through a deft analysis of trends in search words that provide an unfiltered lens into people’s cognition and emotional states as opposed to the projection of a cinematic reel that is a staple on social media platforms like Instagram.
As per the study conducted by Quilt, there has been a 52% increment in online searches related to mental health since the onset of Covid in India in March-April 2020. They also discovered an 80% spike in search terms linked to “suicide” since January 2020.
Their study stresses internalised social codes as a major barrier in the fight to destigmatise mental illnesses and accessibility for relevant help especially at a time like this.
Reading their report, I was struck by the fact that searches such as “feeling alone” and “what to do when you feel lonely” followed a similar pattern as suicide-related searches during the national lockdowns prompted by COVID. Urban loneliness is not merely an offshoot of COVID but this form of social isolation cut open the blinders.
Even before COVID-19, it was estimated that young people (15-24 years) in India suffered from mental health problems. Stress, anxiety, and depression were the most frequently reported. Yet, have access to formal mental health services. A large part of this group is more likely to have greater ease of getting online than getting to a therapist.
I am immunocompromised which means I have had to shelter-in-place for months at a stretch. I too have been shifting between my work screen (laptop), my distraction/interruption screen (phone), and my reward/reinforcement screen (TV). Starting late November, I started stepping out intermittently. I also started watching The Queen’s Gambit. It made me think about a German term that is somewhat known among those who play chess. It is called “Zugzwang”. Zugzwang is an intriguing even if compromising position in strategic board games, particularly those that center turn-based tactics like chess. A player is forced to make a move knowing very well that the move can, inherently, weaken their position. This move signals endgame and is often indicative of two formidable opponents playing each other with very few options left on the table.
Attempts at finding a balance between physical and mental safety during COVID has been a lot like a series of Zugzwang. Except that instead of two equally powerful antagonists, we have a stubborn virus and a largely optionless people.
Friends and clients of mine working retail, healthcare and in other customer-facing environments have been forced to encounter their fears daily right from the beginning. On the flip side, if you are quarantined and deal with something like clinical depression, this means confronting ongoing catastrophic thoughts about prolonged isolation. As the restrictions around physical movement are lifted in most parts of India to a large extent, we are aware that the virus hasn’t disappeared just because we are tired of it. At the same time, a lot of us are also relearning to engage, interact in physical spaces shepherded by caution.
This gets trickier considering that so many conspiracy theorists and anti-maskers have defied guidelines or engaged in toddler-like tantrums leading to substantial roadblocks in preventing the community spread. A lot of discussions around these fears and trepidation occur online.
Chowdhary’s team has been studying these signals through the length of the pandemic. Quilt’s AI uses natural language processing to analyse and cluster texts into themes based on language and content of online posts. They identified 4 key gauges: defining stress, expressing a cry for help, looking for self-relief, expression a desire for connection. The team segregated searches based on publicly sought advice and private search terms. While the public posts center questions around the job, self-esteem, family pressures, personal searches reveal a greater struggle with self-acceptance and confusion around psychological disturbances.
Quilt’s work is quite fascinating in terms of trying to interpret how digital interactions can be used to facilitate better mental health research. Technology’s role in deciphering suicidal ideation online, for example, and pattern recognition cannot be undermined. The quilt took data from 30 Instagram accounts of social media figures who’d died by suicide and programmed its AI to search for patterns in their feeds by studying their posted photographs pixel by pixel. They were hoping to mine something relevant that can help us track signs of mental instability or disturbance online. What they claim to have discovered is starling—an increased presence of a sepia tone— a kind of greyness—closer to the date of demise was common to all the accounts. The color was vanishing from their social media presence. In short, the greyness kept expanding in plain sight but largely went unnoticed.
The other insight they offer is that as opposed to obvious tilts in moods which can be summarized by clear positive or negative remarks, a lot of these feeds displayed neutrality of expression closer to the end of their lives.
Apathy and depersonalization lurked. Angad summarizes this aptly when he says that this indicated that these individuals had “disconnected from the logic of the platform.” In short, a lot of them were showing some kind of withdrawal and no one was paying attention.
Much of the urban life in India is lived online these days and yet so little has been done to discern what exists at the cusp of technology and psychology. Beyond the garden-variety mental health advice available online, a much larger gap exists around online community-building for mental wellbeing. We have an acute shortage of mental health outreach and peer-centered counseling in India that can help address the deficits of our public health system. Mental health practitioners need to be aware of how to translate digital behaviors and offer options for teletherapy.
As I studied the Quilt report, I found it depressing to note that even as digital mental health platforms have mushroomed in the last few years, so much of what they sell is built on ableist slants that serve to widen the distance between illness and wellness. Words like “cure” and “fit” are repeated not-so discreetly in their brand names. The onus on a completely individualized version of wellbeing is somewhat useless during a pandemic that has largely shown there is no easy, singular cure.
Quilt’s study shows us that AI can be harnessed to guide our efforts in understanding mental health stressors outside offline modes of existence. That the myth of “clearly visible sadness” online is exactly that, a myth. That disconnection and disinterest are really the most pertinent way in which we can recognize psychological disturbances. COVID has made it harder for us to ignore these data points and we need to pay critical attention if we are to help people who are possibly going to spend more and more time online in the coming months and years.
(Scherezade Sanchita Siobhan is a clinical psychologist, community catalyst and author.)
(Subscribe to FIT on Telegram)
Subscribe To Our Daily Newsletter Now.