Pretty much what I said, no? You don't use the product and when you do, do it through a filter[1] where you only remember the bugs. Do you use other AI search products and find that they don't show this behavior?
[1] I mean, come on: framing it as "bother not ignoring" is just a dead giveaway that you aren't doing a product evaluation in good faith!
Usually it's non-novel (correctly-harmelessly but unhelpfully, restating the web search results). When it's novel it's because it's wrong.
I would remember the situation where reading the LLM thingy added any value if it ever happened. The weird little UI thing they do where they only show the LLM result if you wait for it to render makes this very easy, I have to scroll up to even consider it.
People are now changing how they search and gather information to use AI. You're automatically discounting that person's experience because they avoid using AI for valid reasons. A lot of times Google AI is outright wrong, not even close to correct. It makes sense to not rely on it and only going back to see if it's improved.
I don't "ignore AI". I just haven't changed how I search and gather information because it's clearly not accurate yet. I still have to fact check it which negates the benefits(time saving). Occasionally I go back to see if it's improved and usually it hasn't. That's not bias.