The problem with AI (specifically LLMs, obviously), is that it redirects traffic from organic human created websites to AI slop, owned by a handful of AI companies. It's ironic that the small web, which Kagi supports, is losing traffic due to AI (including the quick answer feature) with no compensation to the real authors. Furthermore, AI increasingly monopolizes the web, completely contradicting Kagi's "Humanize the web" statement, as there are much higher barriers to entry to become an AI company rather than building a website.
As for the "dangerous precedent", this is unlikely as the reason for not supporting AI is due to moral ground, which Kagi should follow, as it is "humanizing the web". What good is offering privacy when copyrights are ignored and content stolen automatically from small website owners?
As people who are highly interested in search, we would like a say in whether our payments will be spent on developing AI features / AI API, no matter if it is 10000000x cheaper than the classic search, as we do not want any of our money going to AI.
If such a plan is impossible, at least an option that disables all AI "features" that can be easily and accidentally triggered such as typing ? in search, pressing a for assistant, and especially the quick assist button.
