A piece which came through R-bloggers on the subject of AI, caught my attention, and subsequent comment. Of more note, it crystallized a notion which has been floating around my lower brain stem for quite a while. It comes and goes, of course. The notion: is data science, by definition, amoral? Much as laissez faire economics makes no value judgments with regard to fairness or equity of "free market" structures, since it asserts the axiom that, in the end, any market is a fair and balanced arbiter.
Is any application of data/stats/quant disconnected from the use to which it is put? Is it allowable, as I noted in my comment, for the various psych- fields to use data submitted by, or taken unwittingly from, other humans for the sole purpose of directing said humans to behave in ways contrary to their self-interest? Picking which toothpaste, may be who cares. Picking one's government or financial system, may be we should care. Both politicians and marketers have since time immemorial spun narratives to convince those who wouldn't otherwise support their ends, to do so. Thomas Frank's "What's the Matter With Kansas" explored the result a decade ago. The tools of the influence peddlers have only gotten stronger. Vance Packard, in the 1950's, mined the same vein with "The Hidden Persuaders", and that was before transistorized computers even existed.
Is data science just a modern Hessian regiment?