What kind of logical fallacy is X is statistically rare, therefore X is not a problem.

non sequitur. It doesn't follow that because something is statistically rare, it isn't a problem. We don't know if it is a problem or not—too many variables and ambiguity.

The issue seems to be, "Can police brutality be a problem if it only happens infrequently?".  The answer would seem to depend on what one understands as "problem" and "brutality".  It's also important to agree on who it is that experiences the problem.  

If it's a bad thing, it would seem to qualify as a problem for those directly impacted (in the example above, for those brutalized in whatever way by the police).  Whether it's a problem for those who aren't directly involved in the brutality or a problem for the larger society depends on how broadly one defines "problem".

It's easy to run the risk of inviting the ambiguity fallacy when we use (without defining them) terms that could be and are understood in different ways in different contexts.  If we hope to find common ground or agreement among debaters, it's important that each party share a common understanding of the terms being used.