What I said about "wrong information" being dependent on the eye of the beholder -- whether it's an aberration of society or not, is accurate. During the pandemic in 2020, if I would have posted what the Biden Administration concluded -- less than a year later -- about the source of the virus, including using their same exact statements, I would have been blocked on social media, thanks to their expert "fact" checkers. Many who were posting the same exact statements that the Biden Administration later duplicated in their official statements were blocked for disseminating "misinformation" in 2020. Blocked by algorithm, basically.
The facts, or information didn't change at all. But it was the way it was perceived. The same exact information that was deemed unacceptable in 2020 was suddenly deemed acceptable later in 2021.
AI could be used in much the same way, of course. Limit what the internet user sees. We already have a bit of that today, AI, or no AI. I think it's a valid concern, but I don't see AI itself being the problem. It depends on how it's programmed -- just like today's search engine results -- which most people use in research -- all depend on how the algorithm is programmed.
AI does what anyone does: scrape the internet for information, and the way it does it depends on how it is programmed. As for 'duplication', even some news sites just quote other news sites when writing their stories -- which is basically the same thing. When you have a country of 340 million and only ~40K journalists, you're going to get a LOT of duplication and lack of adequate research on stories. AI could indeed make that worse.
It's always up to the individual to check out the sources in the documentation, something most people simply do not do. Instead, they rely on their own confirmation bias. It's human nature. We see plenty of that here on RD, and on social media as well. Even in radio (conservative talk radio being a classic example).
The fact is that most Americans aren't taught to search out documentation, how to do research, or how to use critical thinking. AI isn't going to change that.
Which is why AI can go either way. If it's programmed to seek out diverse documentation, the results may be more accurate. If it's merely programmed to bring up the "top" results from a search engine (which depend on algorithms themselves), the results will probably be skewed.
Ultimately, it's up to the user to use their critical thinking skills.
You mentioned that AI could "pollute the knowledge base". We've already got a lot of that with the major search engines -- sites that probably 90+% of people use to research news and other information. None of the sites seem to bring up the diverse results on the first 3 or 4 pages like they brought up in the 2000's, before the companies decided to monetize the search engines more effectively. There are major items, surveys, and other data that I recall bringing up on the first or second page on one of the biggest search engines in 2005 that I can't even find today, even after searching through tens, if not hundreds, of pages of results. A lot of the time it's page after page after page of the same repetitive crap, rehashed. It's like there has been a limit placed on what search engines will provide in results. Maybe experts in Boolean search terms can get better results -- most of us aren't experts in Boolean search term use. I know I'm not.
So, if search engine algorithms already give most people the same regurgitated pablum as results, how is AI going to make it any worse? After all, AI is just another algorithm, just another computer program -- albeit a very complex one. It's just a tool. It's how the tool is used that matters.
So, I get your concerns, but I think in a way we're already there.
