Honestly, AI-created music in physical spaces has the potential to be even better than what currently exists. Imagine in real time being able to create a soundtrack that matches the mood and vibe of the current atmosphere. When the crowd is bustling, have the audio match that; tone it down when the night wears on, the people have disbursed. Sometimes you go into a Starbucks late at night in Tokyo and they're blasting Led Zeplin–that's probably not exactly what the audience wants to hear. There is potential, with a slightly dystopian tint to it.