Academics, clinicians, students and members of the public came together for the 63rd Maudsley Debate earlier this month to discuss AI and mental health.
Professor James MacCabe presented the motion to the sold-out audience at the Gainsford Lecture Theatre in Denmark Hill, 'The Chatbot will see you now: This house believes AI will do more harm than good to the nation's mental health.' He then touched on how ever more pertinent the question of AI and mental health is, as reports of people turning to AI to support with their mental health increase.
The debate was chaired by Professor Gareth Owen, Professor of Psychological Medicine, Ethics and Law and Honorary Consultant Psychiatrist, South London and Maudsley NHS Foundation Trust. Professor Owen highlighted how this debate brought together not only mental health research and interventions, but also philosophy, technology, our understanding of human/AI relations, and the relationship between clinicians and emerging digital technologies.
Similar to new machines of the past such as the steam engine or the spinning jenny, AI has the potential for huge social transformation. The difference with AI is that these machines have the potential to, quite literally, get inside our heads. Crucial psychological aspects of the AI revolution came through in this excellent Maudsley Debate that grabbled with the question of AI's net benefit or harm for mental health.
Professor Gareth Owen, King's IoPPN
Speaking first for the motion was the psychiatrist and author Dr Iain McGilchrist, joining the debate remotely, who emphasised the risk of isolation when turning to AI, particularly chatbots, and the global impact on mental health this would have. He went on to suggest that a loss of understanding and of creativity due to a reliance on AI, would have further negative impacts on our wellbeing and mental health.
Speaking against the motion, Philosopher Dr Charlotte Blease presented the case that AI could improve access to support and avoid delays both in seeking help and receiving it. Dr Blease also pressed the point that AI could help improve accurate diagnosis and broaden the reach of mental healthcare.
Another Philosopher, Dr Lucy Olser, then came in for the motion, agreeing that there are benefits to AI but that we must understand the harms through understanding its appeal. Dr Olser encouraged the audience to consider that AI is built for profit, and therefore we should question the motivation behind the mental health support.
The main body of the debate was concluded by psychiatrist Professor Argyris Stringaris who posed the idea that how we feel about AI now is how the book was considered many centuries ago. He queried why we're comfortable with some AI use (for example in video calls), but determine other aspects of AI to be bad. He urged the audience to consider whether the field of mental health research and care could be left behind if it does not learn to better negotiate with the changing technological environment.
The audience were then invited to ask questions of the panel, with these touching on vulnerability, stigma, synthetic intimacy, compassion, how to avoid generalisations, the different approaches in primary and secondary care, and questioning whether in the case of not having access to a human being, would having access to AI be better than no access at all?
This debate had so many facets that the speakers approached the question quite differently, which made for a broad ranging discussion, with each member of the panel taking the debate in new, and sometimes unexpected, directions. This is a fascinating topic as the likely impacts of AI are both profound and unpredictable. Perhaps we should revisit this debate in a few years time.
Professor James MacCabe, King's IoPPN
The panellists were then able to sum up their points, with Dr McGilchrist and Dr Osler suggesting that whilst AI presents opportunities in research, it can incubate unhealthy relationships and lacks the wisdom and compassion of a human being. The counter argument from Dr Blease and Professor Stringaris was that whilst badly implemented technology would have a negative impact, well implemented and designed technology would help. They emphasised that turning away from AI wouldn't solve its problems, when it comes to mental health, and it is our duty to use whatever tools we can to reduce suffering.
The recording of this debate, along with recordings of previous Maudsley Debates can be found here.
