I refer to eg this KagiFeedback ticket about inappropriate quick answers.
https://kagifeedback.org/d/10754-inappropriate-suggested-questions
Where Kagi is different than google of course -- is that Kagi doesn't at all force these on the user, you have to ask for it explicitly.
But I was wondering still, maybe in light of strange this legal development it's worth it to get a firm user consent (only needed once) regarding the nature of these things? Like "I understand and agree that these are not real search results but an AI summary" [agree].
I mean it's AI and everyone should be clear on the fact that when you ask for this, it might get some really strange results due to the nature of the technology. Eg Anthropic is not "owned by Jews" but if you ask for an AI summary you get what you order in a way.
I'd hate for someone at Kagi has to sit and manually curate AI craziness around some kind of potential legal liability. Who knows what crazy laws legislators in some jurisdictions may come up with...
CHeers