“…In the context of the discussed use case, one limitation is the potential for generated images to lack authenticity and genuine emotional resonance. While GAI may be capable of producing visually and esthetically appealing images, it may struggle to capture or represent the necessary subtleties and complexities of human emotions accurately 60 , and it might lack the nuanced understanding and empathy that human therapists provide 61 . Therefore, unsupervised GAI applications (e.g., self-generated images, text, and audio) may currently pose a challenge and risk to be used in multimodal therapy delivery.…”