Noticed an interesting article about using computer software to decipher an ancient language, in this case Ugaritic. The article is here. There are, of course, a few caveats which are worthy of note. Ugaritic was chosen because it is well understood so the computer’s results could be checked. Furthermore, it only works if a series of rather special prerequisites are met: the unknown language must be a reasonably close cognate with a known language, its writing must map onto that of the known language, and others. It also looks as though it only works for alphabetic scripts.
Still, perhaps Google will provide a Ugaritic language option soon!