I think the detection and logging aspect has a lot of promise as a cheap and widely deployed system.
I'd worry about the liability from false positives with a active response system. Even something as limited as locking the doors to a room to contain the shooter could lead to liability from someone being trapped with the shooter.
If you assume a large indoor setting, something like a mall or school, maybe combine the detection network with a evacuation guidance system that leads people away from the incident. Mass shootings are pretty rare so maybe you could use it as a general "smart evacuation" system. OTOH there's probably a crowd psychology problem in there that makes that unfeasible.
Good point about false positives. The device could upload images to a human operator who would activate the defenses. This would add some reaction time, but it would reduce the potential for false positives.
If the device has image recognition to detect firearms being brandished, the operator could trigger the defenses before the first shot is fired perhaps.
I'd worry about the liability from false positives with a active response system. Even something as limited as locking the doors to a room to contain the shooter could lead to liability from someone being trapped with the shooter.
If you assume a large indoor setting, something like a mall or school, maybe combine the detection network with a evacuation guidance system that leads people away from the incident. Mass shootings are pretty rare so maybe you could use it as a general "smart evacuation" system. OTOH there's probably a crowd psychology problem in there that makes that unfeasible.