You know how Google’s new feature called AI Overviews is prone to spitting out wildly incorrect answers to search queries? In one instance, AI Overviews told a user to use glue on pizza to make sure the cheese won’t slide off (pssst…please don’t do this.)
Well, according to an interview at The Vergewith Google CEO Sundar Pichai published earlier this week, just before criticism of the outputs really took off, these “hallucinations” are an “inherent feature” of AI large language models (LLM), which is what drives AI Overviews, and this feature “is still an unsolved problem.”
Never understood this argument. It has nice rethorics but proves nothing. the problem with text prediction != intelligence is that we can’t even define intelligence. I hate AI as much as the next guy here, but that doesn’t mean we should do lazy assertions and quoting dumb arguments.
You’re not wrong that we don’t have great definitions of intelligence, but that doesn’t detract at all from criticism against these LLM’s being pushed on us and the massive amount of praise and hype being directed at what’s essentially advertising platforms.
All you need to understand is that AI models are making the internet a lot worse, and it’s going to get worse and worse and people are not happy about it. The exact details and semantics about the wording is almost irrelevant in the face of our need to collectively demand better from our tech companies and not act like these LLM’s are real entities that deserve respect.