It hallucinates so frequently, that I've found it unusable. Just ask it to provide sources for all of its claims, then watch how many times it reverses its original answer. It's mind-boggling to me. I've repeatedly tried to give Google AI Search a chance, but after weeks of seeing this behavior I can't trust it anymore.

Imagine winning search by being the fastest and most accurate, then replacing it w/ constant hallucinated errors.

For me, I use it maybe ones or twice a month when my traditional search on DuckDuckGo fails. And, anecdotally, it is overwhelmingly wrong, and worse, confidently so.

Given my sample is biased as I am using it for an outlier of difficult web searches (including web searches localized to Iceland). But I would think that if your parent was correct, at least majority of my uses would be correct, but in my experience, Gemini fails most of the the time traditional web search fails. So to me at least, it is effectively useless.

For people who use it in place of normal web search, it is probably worse then useless, because of how confidently wrong it is.

The argument is not that Google's AI summaries are infallible.

The argument is that there's so much garbage on the open web that clicking through to the source has a lower signal-to-noise ratio than the AI summary.

This rings very true in my experience.

Google's AI summaries definitely make mistakes, but in most cases they're higher signal than the source websites.

I think they model the model they use for these, probably at more cost to them. They've been much better as of late, imo