AI Helps Track Kids' Feelings During Horse Therapy

In the Kimberley region of Western Australia, a program called Yawardani Jan-ga brings horses and Aboriginal youth together. The focus is on social and emotional well-being. Children take part in therapy sessions with horses. Helpers watch closely and write down what they see.

Today, most of this tracking happens by hand. Workers use checklists, run interviews, and keep progress notes. This process is slow. It also depends on personal judgment. One helper might notice something that another helper misses. There is no shared standard across different sites or staff members.

A research team created a digital tool to help. It uses artificial intelligence to study photos from the sessions. The tool comes in three pieces. One piece identifies nine therapy activities using a vision-language model. Another piece looks at physical contact between the child and the horse. A third piece reads body posture to spot confidence-related states and explains its reasoning.

The system was built and tested with real images gathered during actual therapy sessions. Instead of only relying on human notes, the AI can review what happened between the child and the animal. This may free up time for workers and give a clearer picture of each child's progress.