When search results are just links to content related to a topic, the provider isn't responsible for the accuracy of the content, which is created by others.

Maybe google should try to make a service that just shows links to useful pages to answer the search requests.

Which is such an amazing detail, and really is the clincher.

They would have to revert their AI “answers” until they can deliver consistent accuracy on their answers. Which isn’t even a possibility with AI.

I hope someone brings this to the US as well.

When search results are just links to content related to a topic, the provider isn’t responsible for the accuracy of the content, which is created by others. But they are responsible for their own summaries and other provider-created content.

This is clearly a reasonable line to draw. This is not content created by others. It's your robot. Fix your shit.

Saying "I slapped a disclaimer on my libelous robot that says that it may generate libel" doesn't grant you the right to be libelous.

This is entirely reasonable and I hope this understanding catches on in courts worldwide.

If I go make my own summary/parody/spinoff/reaction video based on someone else's content, I'm responsible for the media I created. Simple as that. Same should apply to companies.

This is where the "nobody needs AI to search the internet" part comes from (emphasis mine):

Historically, any potentially harmful content surfaced by search engines has been protected from direct liability because that surfacing was considered largely unavoidable when helping users sort through an enormous tangle of information online. But the German court emphasized that AI search engines do not enjoy those same protections because AI summaries merely provide “an additional function—one without which the use of the search engine would still be (and is) possible, and without which users are perfectly capable of finding results amidst the ‘flood of data.’”