3 ms·
Beyond Bias and Compliance: Plurality in AI Ethics
- version_five 3y agoI didn't think this was very good - it was barely coherent. But it makes a very good point: The fixation on bias and compliance within the AI ethics market leaves out actual ethics I'd add that all ML is stereotyping, that's literally what it does, it's fitting a curve, not reasoning. So any application where stereotyping is inappropriate shouldn't use ML.
- dragonwriter 3y agoI disagree that that is a good point, because I've seen lots of discussion of other ethics issues from the same people th as t also highlight bias, and bias is, in many domains of—current, not merely potential—AI application an issue of significant concrete ethical relevance.