DePoe...?
Now, imagine the sheer quantity of programming required for a computer to create such a clue. There would have to be a database of American phonemes and variants (depot, depoe, etc.) There would have to be a gazetteer of famous people, and it would have to be huge. There would have to be a choice mechanism: is it pronounced dehpo or deepo, with dehpo being disqualified for the purposes of creating this clue, which mechanism, privately, I believe to be impossible to create short of including every possible pronunciation of every combination of letters.
There would have to be a means of compiling idiocies (i.e. arbitrary nonstandard usages and pronunciations for the purpose of wordplay) and some sort of selection mechanism for same. These factors, and others I haven't mentioned, imply a complexity in the programming of wordplay that is, as a guess, hundreds of orders of magnitude greater than that of chess.
Not to denigrate chess, not at all. However, fact of the matter is that chess CAN be analysed effectively completely by exhaustion of possibilities, whereas language cannot, not even close. Information is constantly entering the world of language; witness the number of neologisms created weekly, right? Not so chess.
FReegards to you!