No, not really -- that's why it was originally purposed as an encryption algorithm.
Providing two like data streams will produce the same hash, allowing one to confirm the validity of one file against another to a very high probability, however you can't take a hash and "reverse engineer" it to produce any usable information.
For example, any file that's run though the algorithm to produce a hash will produce a has of the same size (16 bytes), so the MD5 hash doesn't even provide any info on what SIZE the original file.