Flutterby™! : Errors are a feature

RT Arri Blais :verified_trans: @arri@hachyderm.io

May be an unpopular take, but I think the main risk of ML is not that it's good at what it does.

It's actually very bad, worse than humans by far at what it does. But because people collectively believe it is good at stuff, people are focusing so much on "the singularity" or "it replacing [x job]" the real risk continues unhindered.

This is a good example of the true goal of ML: introduce bias and inefficiency in areas where inefficiency is desirable to capitalism.

UnitedHealth uses AI model with 90% error rate to deny care, lawsuit alleges — For the largest health insurer in the US, AI's error rate is like a feature, not a bug.