Whitepaper: What can AI offer to the field of mental health?
As noted in our previous blog What are the challenges of digital mental health care solutions?, there is a growing and unmatched demand for mental health care worldwide. Digital solutions can significantly help address this problem but they can’t provide all aspects of interaction needed for good treatment. In this blog we look more closely on how artificial intelligence (AI) can provide the next crucial development step for these digitized mental health care solutions. For a more in-depth look, see theWhite paper by our Formulator team’s AI guru Giulio Jacucci.
AI is good at scaling
AI can be very effective at solving complex problems on a large scale. Recent developments such as ChatGPT have showed that machine learning based language models and chat bots can quite reassuringly act as conversational agents and interact with users in discussions that uncannily mimic meaningfulness. Instead of feeling intimidated by this, we should look at the opportunities for integrating AI as a part of good interaction between health care providers and clients.
Research over the last decades has also shown that digital interventions as substitutes for mental health professionals are useful, but not as effective as combining digital intervention with face-to-face support. We propose that digital solutions should not try to replace human interaction in mental health interventions but rather enhance the healing power of a good therapeutic alliance. All therapeutic work starts by creating a trusting environment. A therapeutic alliance, that consists of bond, goal setting and tasks, has been found to be one of the most important aspects of successful therapy regardless of therapeutic school or education. Formulator facilitates the meaningful formulation of goals and tasks.
Enhancement, not replacement
In his paper Giulio describes three potential uses for AI in mental health care:
1) Sensing, detecting and predicting different diagnoses and disorders,
2) Decision support in choosing the best intervention and
3) Carrying out the actual intervention.
With Formulator we propose an integrative approach: systematically developing hypotheses about the causes, precipitants, and maintaining influences of a person’s psychological, interpersonal, and behavioral problems, as well as a plan to address these problems. Although this case formulation process is complex, it can be systematized and thus performed effectively by AI.
The formulation sets the stage for a good therapeutic alliance and interventions. Imagine both the patient and professional having a shared starting point that allows them to dive straight away into the most human part of the therapeutic process - interaction.
Combining rigor and intuition
Although mental health care and psychotherapy are often associated with a certain reliance on intuition and empathy, they are based on theoretical models and often quite rigorous techniques. Still even trained professionals may occasionally find it hard to work in a highly systematic manner and often the meaningful moments in a mental health care situation can arise from a small signal or moment of intuitive insight. The potential of AI enhanced therapeutic interaction comes from AIs capability for complex rigorous analysis and the genuine warmth of human interaction.
Inviting AI to the conversation
We are only starting to see the potential of AI in healthcare. So far most use cases have been in classifying and categorizing structured data. As language models evolve further, we will see new developments in the use of unstructured data, such as conversations. The ethical thing is to involve AI as a part of the interactions and conversations, to use it as an adviser for enhancing our own professional skills.
Are you interested in learning more about the benefits and challenges of AI solutions? Download our whitepaper “What can AI offer to the field of mental health?” to have a deep dive into our approach. We explain how we see that combining psychology and technology should look like, and what are also the challenges when doing it.