alt Hacker NewsMost people Google things they're unfamiliar with, and whatever the AI Overview generates will seem reasonable to someone who doesn't know better. But they are wrong a lot.
It's not just the occasional major miss, like this submission's example, or the recommendation to put glue on a pizza. I highly recommend Googling a few specific topics you know well. Read each overview entirely and see how many often it gets something wrong. For me, only 1 of 5 overviews didn't have at least 1 significant error. The plural of "anecdote" is not "data," but it was enough for me to install a Firefox extension that blocks them.
Replies
I've found that the AI Overview is more accurate than it used to be... which makes it much worse in practice. It used to be wrong often enough, and obviously enough, that it was easy to ignore. Now it's often right and usually plausible, which makes it very tempting to rely on.
Here's my paraphrase of the best description for "seems reasonable" AI misinformation that I've yet seen. I wish I could credit where I first heard it:
AI summaries are akin to generalist podcasts, or YouTube video essayists, taking on a technical or niche topic. They present with such polish and confidence that they seem like they must be at least mostly correct. Then you hear them present or discuss a topic you have expertise in, and they are frustratingly bad. Sometimes wrong, but always at least deficient. The polish and confidence is inappropriately boosting the "correctness signal" to anyone without a depth of knowledge.
Then you consider that 90% of people have not developed sophisticated knowledge about 90% of topics (myself included), and it begins to feel a bit grim.
I found it is very accurate for legacy static-web content. E.g. if you ask something that could easily be answered by looking at wikipedia or which has been answered in blogs, it will usually be right.
But for anything dynamic (i.e. all of social media), it is very easy for the AI overview to screw up. Especially once it has to make relational connections between things.
In general people expect too much here. Google AI overview is in no way better than Claude, Grok or ChatGPT with web search. In fact it is inferior in many ways. If you look for the kind of information which LLMs really excel at, there's no need to go to Google. And if you're not, then you'll also be better off with the others. This whole thing only exists because google is seeing OpenAI eat into its information search monopoly.