Music has long been recognized for its healing powers, and when combined with artificial intelligence, this potential becomes even more precise and accessible. This sentiment, shared by Li Xiaonuo, vice-dean of the Research Institute of Music and Art at the Shanghai Conservatory of Music, resonated throughout the recent World Artificial Intelligence Conference in Shanghai.

Among the event's many technological marvels, the globally premiered AI music therapy pod from the Shanghai Conservatory of Music stood out, offering a serene, personalized healing experience.

Inside the pod, users are equipped with monitoring devices, such as EEG(electroencephalogram) headsets, HRV (heart rate variability) sensors, and conductive nail monitors. Reclining in zero-gravity chairs, they observe a large screen displaying real-time EEG data alongside vital signs, such as heart rate and blood oxygen levels. Meanwhile, ambient music featuring the natural sounds of chirping birds or flowing streams envelops them, providing a peaceful escape from the outside world.

Ren Shihong, a postdoctoral researcher in artificial intelligence music generation at the conservatory, explains that the experience begins with a personalized questionnaire. Users provide details about their health status and birth date, which, combined with real-time physiological data, drive an AI music generation model to create a multisensory experience complete with synchronized visuals.

After the session, users receive a dataset of their therapy process and a personalized music prescription. "The system enhances attention, aids in healing and relaxation, and stabilizes heart rate," says Liu Hao, director of the AIMT Lab at the conservatory.

Li says: "The music therapy pod is a crystallization of the integration of technology and art." The college's music aesthetic psychology experimental center, established in 2017 and now part of the AIMT Lab, has published numerous Scientific Citation Index papers, laying the groundwork for the therapy pod.

Research into converting brainwaves to sound began in the late 20th century. Each person's resting brainwaves are unique and can be transformed into sound waves through algorithms, Li explains. AI's role in this field is crucial, as it enables the rapid retrieval of personal trait patterns and generates soundscapes based on real-time physiological indicators.

"People's responses to sound have commonalities and individual differences, involving high-level evaluative memory, conditioned reflex memory, and personal musical-cultural experiences — all part of high-level music cognition," Li points out, emphasizing the need for massive data computing, which AI provides. The goal is to leverage AI's power to reach even quantum-level computing.