definitely. also once training gets easier anyone can train on a set of accepted laws and rules and run their own (this is true open source ai, ai which you trained the weights yourself). This is basically a form of *automated reasoning* which can help us understand and point out when policies go awry.