Here are some related online sources, which I found extremely insightful:
Two blog posts about Goodhart’s law by David Manheim. He appears to be one of the first to notice, or at least clearly state, that the law applies quite similarly to artificial intelligence, engineering and complex social systems. The first post discusses the problems with ‘measures’ in complex systems and the second elaborates on the relation of such measures and the actual goals.
An aeon blog post by Scott E Page explainging why diversity often trumps excellence in complex systems (hint: excellence by what measure?).