Because majority of people don't understand how LLMs work, what hallucinations are, or the capabilities of AI tech. Like lawyers https://www.abc.net.au/news/2023-06-24/us-lawyer-uses-chatgp... airline website devs https://bc.ctvnews.ca/air-canada-s-chatbot-gave-a-b-c-man-th... law enforcement https://gizmodo.com/deepfake-cheer-mom-sues-for-defamation-1...

No, people in general cannot understand changes in technology.

GPT3.5 used to hallucinate a lot but it was drastically improved with 4, to a point where I can't even remember the last time it did it. It feels like people are exaggerating or basing it on the experience with old models.