The more sophisticated AI models get, the more likely they are to lie by Jacek Krywko from Ars Technica - All content on 2024-10-04 19:39 (#6R80C) Human feedback training may incentivize providing any answer-even wrong ones.