That's hilarious, I was working on https://pomax.github.io/music-theory-js/public/ over the weekend (working on adding a chord arranger to complement the cell arranger, so you can write a C3-root I–V–vi–IV as literally a "C3 I" cell with some duration, followed by just dropping in V, vi, and IV cells with some duration.
Short of a genuinely complicated ML system which having been fed a bunch of (say) Jazz Pianists midi data for chord patterns etc, I think you could probably do it with some kind of weighted (let's say by the user) Markov chain/state machine with some kind of graph for which chord voicing should be used relative to those around it: e.g. "This Dm7 has a F in it, the next chord is a Cmaj7#11 so move the F up a semitone[Within the same octave]"
The above system should be able to come up with coherent movement of voicings (and fingers...) although I can't imagien that it would sound very human.