To recapitulate, computer science grad students developed a pattern recognition software that took pre-established patterns developed by biblical criticism scholars and ran the Bible through it. Surprise, the results matched the previous scholarly analysis.
Here is the problem: if the consensus around the criteria developed by the biblical criticism school is wrong, this would not prove it. All we proved was "yes, you can find distinct patterns in the fashion identified." Whoo hooo.
For example, suppose I wished to identify who "really" wrote Shakespeare's works. I start with the supposition that someone who writes plays can't _also_ write poetry, so clearly the author of the "Sonnets" (the "S" source) is different from the "Folios" (the "F"). I explain the fact that we find poetry in the Folios as the work of a redactor who took various pieces of poetry, some of which are similar to that in the S, although some of the poems are different. In addition, scholars note that even within the Folios, we have several different authors. Some scripts are profoundly influenced by Hollinshed, and are referred to as the "F(H)." Certain plays have mythological elements, F(M), and some scholars even postulate multiple F(M) sources.
Then, based on several centuries of such analysis, others develop a computer program that looks for the patterns identified by scholars. Amazingly, they discover that the works attributed to "Shakespeare" break down in almost exactly the way scholars describe.
It is interesting, but it is not proof.