Honestly, I'd be more surprised if it couldn't. That different populations might have statistical skeletal "markers" isn't very surprising.
In fact, I'd be less surprised to hear that different ethnic groups have different skeletal features than the only thing the AI is going on is the shadow of melanin in an x-ray (though I could also believe that).
In fact, I'd be less surprised to hear that different ethnic groups have different skeletal features than the only thing the AI is going on is the shadow of melanin in an x-ray (though I could also believe that).