Random thoughts: 1) My first thought would be that such a conversational search UI would never hit the original website, thus drying up monetization for the co

Random thoughts: 1) My first thought would be that such a conversational search UI would never hit the original website, thus drying up monetization for the content creator. Kind of like snippets in current search. Then I realized that there is no "original website", no singular source for the AI answer given. That's even worse. Search is the main traffic controller of the web and now has zero reasons to send any traffic your way. Not only does it not have any reason to send traffic, it simply can't. You can't back trace an AI answer to a (single) source. 2) The implication of the above point may be that for human creators, producing new content for the open web becomes (somewhat) pointless, even in the case where you do not even care to monetize it. An exception is video, for now. You should not expect this change to lead to a drying up of content for AI though. Google/Microsoft will simply resort to making deals with major content platforms. That's where most content is produced anyway. 3) An interesting new problem for Google/Microsoft is the increased accountability regarding correctness. Right now, search results do not have to be correct. You're just showing results produced by other entities. This changes when you confidentially give an answer directly in the search UI though. Now you own the answer, even if this expectation is technically inaccurate. Since GPT3 is often confidentially incorrect, that's a major challenge. And we should hope that this problem remains for a long time, because the moment GPT3 is extremely accurate and correct, we're in a different world entirely. 4) Safety and political correctness is another concern. Google/Microsoft will likely deliver AI that is a little too safe instead of not safe enough. The implication is that instead of being a controller of traffic, it becomes a controller of opinion. Of right think and wrong think. Which may be ever changing and subject to institutional pressure. Combined with point 3, you're no longer a proxy. You're the source of truth and the source of what's right. 5) Some say that the open web is already close to dead thus the already ongoing trend of people resorting into private spaces (Discord, Mastodon, etc) might accelerate. An open web and a closed human web. Even in the closed human web AI-generated content can and will penetrate so there will be attempts to detect it. Which is likely futile, because one can use AI-generated content and then tweak it to make it more human-like. 6) We need legislation ASAP. None of our regulations and laws are designed for what is coming. Do you really want a congressional hearing 10 years from now where a 75-year old asks "So this AI thing, is it like a smart calculator?", after all damage has already been done? 7) We should unite in at least some basic pushback. Stop thinking in individual use cases like AI art, Copilot and GPT3. Think about the broader problem where all of human creation is sucked up for free and without permission, after which all benefits are centralized up to the point of even replacing you. It's a shared, human problem. We're unlikely to stop it but at least don't make it this easy.