Potential application of AI in psychotherapy
A hypothetical scenario of a human patient consulting a robot psychotherapist Photo: TUCHONG
Traditionally, psychologists tend to consider subtle fluctuations of emotion, complex human cognition, and underlying motives for behavior to be domains that machines can’t fully grasp. It was once believed that psychotherapists could never be replaced by machines, as their work entails an in-depth understanding of the emotions and experiences unique to humans. However, with the rapid advancement of technology, such views have been challenged. It is now necessary to re-examine the future status of the profession of psychotherapy in society and the role of machines therein.
Evolution of AI
Early AI classified data according to predefined rules without being able to understand or establish underlying connections between data. With the emergence of machine learning, especially deep learning, large language models (LLMs) have made tremendous progress in this regard. These models have grasped not only the surface structure of language, but also the context and socio-cultural background, enabling them to emulate human thought processes to some degree. This has propelled AI to new levels in terms of understanding and interacting with humans, opening up new opportunities for the deployment of AI in the field of psychology.
Construction enables understanding
LLMs are capable of constructing discourse along multiple dimensions, which involves not only semantic accuracy but also contextual adaptation and emotional association. This ability enables a new form of “understanding.” At the technical level, this new understanding is built upon complex algorithms and large-scale data training. At the functional level, machines can largely understand and respond to human emotions through language and context construction. While their “understanding” does not derive from inner emotional experience, machines are able to reproduce the external form of human understanding to a large extent and manifest adaptability and coherence when offering support and advice. This has allowed for the application of machines in psychotherapy and other areas that entail in-depth understanding and empathy.
Understanding enables healing
Understanding is deemed crucial to the healing process by numerous schools of psychotherapy. In psychoanalysis, understanding constitutes the basis for discerning the inner conflicts and subconscious desires of the patient. In cognitive behavioral therapy, understanding serves to identify and correct cognitive distortions. In person-centered therapy, understanding is embodied in profound empathy towards and unconditional positive regard to the patient. Current AI systems, with enhanced comprehension, are capable of emulating such processes in which understanding serves as a driving force that can guide individuals to self-discovery, self-understanding and ultimately self-healing.
In psychoanalysis, while AIs cannot intuitively sense the conflicts in the patient’s subconscious mind, they are able to understand the psychological motives behind the patient’s behavior by analyzing language features. In cognitive behavioral therapy, they can understand the connections between events, perceptions, and emotions and reveal negative thought patterns based on logical reasoning and data, thereby leading individuals to recognize cognitive distortions and adopt more positive ways of thinking. In person-centered therapy, AIs can emulate unconditional positive regard through verbal feedback, helping individuals enhance self-identity and self-worth. They can also provide coherent and consistent responses while avoiding certain human biases in therapeutic processes, which potentially benefits many patients, as stable relationships and reliable environments are known to improve therapeutic effects.
Restructuring of the industry
The application of AI can considerably enhance the accessibility, cost-effectiveness, and convenience of mental health services, making such services readily available to a larger population, particularly groups for which access to professional mental health services is difficult. This contributes to improving the public understanding of mental health issues, which in turn increases demand for mental health services.
Human psychotherapists may need to collaborate with machines, delegating standardized or routine counseling tasks to AIs while reserving more complex cases for human experts. While such collaboration will allow psychotherapists to increase efficiency and provide more comprehensive and diversified services for their clients, it may also lead to intensified market competition and changes in professional roles. Human psychotherapists may face reduced job opportunities or potential replacement by AI, necessitating the acquisition of new skills for collaborating with AI or specializing in specific fields of psychotherapy. The coexistence of human and AI psychotherapists is likely to become an important area of research.
From a market perspective, the introduction of AI may drive technological innovation and the diversification of service models in the mental health industry. This could lead to an increase in AI-empowered mental health applications and platforms, as well as personalized and customized services. These innovations could attract new investors and practitioners to the industry, thus accelerating its overall development.
Yan Wenjing is an associate professor from the School of Psychiatry at Wenzhou Medical University.