How often have you heard that Artificial Intelligence (AI) will not be able to demonstrate empathy? That empathy will be the Achilles Heel of AI? Let me break the good news to you: Generative AI (Gen AI) will get there faster than we think. In an absorbing paper called Situational Awareness, The Decade Ahead, Leopold Aschenbrenner, who was part of OpenAI’s superalignment team until recently, says, “If there is one lesson we’ve learned from the past decade of AI, it’s that you should never bet against deep learning.” First off, this means deep learning will break barriers faster than we can imagine and improve the empathy quotient of AI. By implication, AI can potentially advance mental health care, even revolutionise it. Mental health care remains an area of life that requires greater attention than it currently has. Second, remember to read Aschenbrenner’s paper. It is packed with ideas and information that will help you better frame the progress of Gen AI.
But let’s get back to Gen AI and examine the size of the challenge we face with mental health. In 2022, the World Health Organisation’s (Who) World Mental Health Report—Transforming Mental Health for All said that by 2019, almost a billion people were living with mental disorders. This included 14 per cent of the world’s adolescents. Suicides accounted for one in 100 deaths and 58 per cent of suicides occurred before the age of 50. It further pointed out that mental disorders are the leading cause of disability, responsible for one in six years lived with disability (depressive disorders are the single most significant contributors to non-fatal health loss). According to a scientific brief released by Who, depression and anxiety went up by more than 25 per cent in the first year of the pandemic.
The resources to address mental health are hopelessly inadequate. Who’s Mental Health Atlas for 2020 showed that a mere 2.1 per cent of government expenditures worldwide were on health. Globally, the median number of mental health workers is 13 per 100,000 population. Can Gen AI provide the interventions required to heal us from the wounds of mental distress?
In an article called Emotion AI, explained published by MIT Sloan, Meredith Somers, who covers behaviour and policy sciences, explained how Emotion AI, also known as Affective Computing, was being used to measure, comprehend, simulate and respond to human emotions. This branch of AI was being used “to improve everything, from marketing campaigns to health care.” In fact, researchers at MIT Media Labs had built an algorithm using phone data and wearable devices to predict depression. The next step? Algorithms that help manage depression.
The opportunity is so obvious that one recent scientific paper on evidence-based digital health research says an estimated 20,000 mental health apps are already available. The market for mental health applications is expected to be USD 17.5 billion by 2030, growing at a CAGR of 15.2 per cent from 2024 to 2030.
The reason for this boom is clear. Using AI is more convenient—and cheaper—than looking for difficult-to-find healthcare professionals. Woebot Health, a leading therapeutic chatbot company, surveyed adults in 2021 and found that 22 per cent had already used a mental health chatbot and 47 per cent said they would be interested in using one if needed. Of those who had used the chatbot during the pandemic, 44 per cent said they had not seen a human therapist.
The apps are turning out to be effective, too. Flow, used to manage depression using a brain stimulation headset and therapy, saw 81 per cent of its users feeling better in three weeks, 34 per cent reported an improvement in mood and 29 per cent reported a reduction in suicidal thoughts. The app reported an increase in sales of around 247 per cent in 2023.
In a webinar in May 2024, Gartner’s VP Analyst Mary Mesaglio explained the growing popularity of mental health apps and their implications for business. She spoke about a Chinese AI bot called Xiaoice (“Always there”) that was providing comfort and counselling to 660 million users worldwide. Xiaoice is designed to create emotional bonds with its users. Mesaglio, of course, used this example to explain how, in a business context, the marketing game will move from attention to intimacy.
We know that the most obvious mountain for AI to climb is human empathy. Empathy involves possessing a degree of curiosity in understanding and appreciating the other person’s point of view so that one can behave compassionately. It is becoming even more important to put the might of AI behind empathy because the trait appears to be drying up in the hearts of humankind. We hear dismal anecdotes of its decline among school teachers, government officials, in the workplace, with law-and-order officers, etc.
Signs are emerging around us of how important it is to supplement human empathy. For instance, a service called Empathy uses AI tools to help users manage and overcome bereavement and grief. The platform has helped some 40 million people regain stability by focusing on compassionate care. The company had raised USD 90 million in funding and was valued last March at USD 400 million. We may view this with skepticism today, but it is tomorrow's reality: Gen AI will firmly be in the empathy game.