Home Cyberpsychology & Technology AI Holds Promise for Enhancing Art Therapy, But Requires Further Refinement and Human Oversight

AI Holds Promise for Enhancing Art Therapy, But Requires Further Refinement and Human Oversight

Published: Last updated:
Reading Time: 4 minutes

Artificial intelligence (AI) is increasingly being utilised in various fields, and its application in mental health care is a burgeoning area of interest. One promising development is the use of AI in art therapy for individuals with mental disorders. This innovative approach combines the creative process of art therapy with the analytical capabilities of AI to provide new avenues for expression and treatment. The findings were published in Psychreg Journal of Psychology.

According to the National Institute of Mental Health, 1 in 5 adults in the US suffers from a mental disorder. These disorders can lead to significant distress and social isolation. Traditional art therapy has long been used to help individuals reflect on their emotions, boost self-esteem, and reduce social isolation. However, the integration of AI into this therapeutic process offers new possibilities.

Mais Aljunaidy, PhD from Texas State University said: “The motivation was to study the potential of using new and emerging technology in the field of mental health. The study aimed to assess to what extent AI can ‘actually’ understand human emotions and reflect them in art creation, which is not fully determined. The study also aimed to discuss how AI was used in mental disorder art therapy and to assess how AI interprets the three most common mental disorders (depression, bipolar, and anxiety) through visual art and compare it to interpretations from a human point of view.”

AI tools can collect and analyse vast amounts of data, enabling researchers to explore how art therapy can be optimised for individuals with mental disorders. For example, the University of Waterloo has developed a digital art therapy programme called DeepThInk, which allows users to create digital paintings that vividly express their emotions. Another example is CareYaYa, an AI art therapy tool that has shown promising results in supporting individuals with dementia.

Despite its potential, AI in art therapy is not without its limitations. Aljunaidy said: “AI is not completely perfect in in-depth interpretation and analysis of mental disorders’ symptoms as it understands mental disorders’ symptoms literally. The extent AI can ‘actually’ understand human emotions and reflect them in art creation is not fully determined.”

One major concern is the extent to which AI can truly understand and interpret human emotions. AI systems can generate art based on user input, but these creations often reflect a literal interpretation of emotions rather than a nuanced understanding. For instance, AI-generated images depicting mental disorders frequently use dark colours and disturbing imagery, highlighting despair and distress.

To address this, researchers have experimented with altering the input given to AI tools. By using more optimistic language, they were able to generate less distressing images. For example, when the term “depression feelings art in an optimistic way” was used, the resulting images were more colourful and less negative. This approach highlights the need for continuous refinement of AI algorithms to better capture the complexities of human emotions.

A comparative study explored the differences between human and AI interpretations of emotions associated with depression, bipolar disorder, and anxiety. Human-created images tended to capture the depth of emotions in a more abstract and indirect manner. For example, a depiction of depression might feature tears surrounding a heart, symbolising deep sadness. In contrast, AI-generated images often took a more straightforward and literal approach.

Despite these differences, AI tools have shown potential in providing remote assistance and continuous support for art therapy. They can analyse and recognise patterns in user input, offering a level of image analysis and recognition that can complement the work of human therapists. In some cases, AI has matched human experts in analysing cognitive and emotional states, such as in a South Korean study where computer technology assessed the level of concentration in elderly individuals using a structured mandala art therapy tool.

Aljunaidy also notes the implications of these findings: “Although AI tools can significantly support the healthcare system and ease the pressure on it, they might not be completely reliable in art therapy for mental disorders and still benefit from a lot of improvements. Having a human filter or oversight to mitigate the literal interpretation of AI to patients’ emotions is still needed.”

The use of AI in art therapy also raises important sociocultural considerations. Not everyone is aware of or comfortable with using AI-powered mental health applications. Privacy concerns, stigma surrounding mental health, and a lack of access to technology in low-income regions are significant barriers. Additionally, AI lacks the human qualities of empathy, compassion, and trust, which are crucial in mental health care.

To address these challenges, it is essential to educate the public about the availability and benefits of AI mental health tools. Ensuring data privacy and providing clear consent mechanisms can also help build trust. Furthermore, AI tools need to be made more accessible and affordable to ensure they can benefit individuals in diverse socioeconomic contexts.

Looking ahead, Mohamad Nadim Adi, PhD, the first author of the study, outlines the next steps: “The future plans are to incorporate generative AI as an active partner in mental health therapy. The ideal AI tools should have genuine human feelings and reactions. Whether this is achievable or not is unknown. In the meantime, researchers should continue to refine the AI algorithms to have a better reading and interpretation of human feelings and emotions, especially when it comes to art therapy for mental disorders, where situations like being at risk might occur and need immediate assistance.”

© Copyright 2014–2034 Psychreg Ltd