When you ask an AI bot for an image of the Founding Fathers or a group of German soldiers from 1943, you expect… something. You probably don’t expect what Google Gemini has been delivering, which is a set of images that goes heavy on diversity and light on historical accuracy. And when you ask ChatGPT a question, you probably don’t expect total gibberish in response. It’s been a very strange couple of days in AI land, but it makes us wonder: what do we actually want from AI?
Gemini, ChatGPT, and the other AI bots are out of control – The Verge
Gemini was either built incompetently, or deliberately tuned to generate nonsense, in which case Google is in the propaganda business. Either way, can we trust AI systems? No. Can we ever trust Google? This is not the first time they’ve screwed up big time – and then appeared to have lied about it.
