'Junk DNA' is any DNA whose purpose was unknown when the article / book was written. But to return to your question, not necessarily.

First, we are usually concerned with the (dis)advantages of mutations when they occur in coding regions, which are definitely not junk DNA.

Second, just because a sequence does not encode any useful information does not mean it is useless. For example, it could be holding a coding region away from another, so both can be transcribed at the same time. Or it could be structurally important in the way the chromosome is folded.