Ah, I see the thread has grown - and now I see what the problem is. You have a fundamental misunderstanding of what entanglement is, I'm afraid. The famous Bell theorem certainly rules out any local, causal, definite model. Non-locality applies even where something at A must cause something at B, so that would mean information moving faster than light even in ordinary EPR experiments - unless our whole idea of things interacting locally is utterly wrong. Likewise causality is not really optional unless you are happy with time-travelling waves (The TI) or such-like. So "crossing the event horizon" is no different from the "faster-than-light" interpretations of ordinary entanglement. In fact that's exactly what it is. But in a no-collapse interpretation - see my other reply - such as MWI, there is no need for any information to be exchanges at all. The observed correlations are just the effect of including the macroscopic observer in the system. One again, using a BH as the observer makes not a blind bit of difference.