As my seven or nine regular readers might imagine, I now have something of a backlog of articles about AI and machine learning, as, unlike a large language model, I’m not ingesting hundreds of articles a second and spitting out my pattern-recognized conclusions.
So, with the news in the past week of AI companies buying tons of physical books to scan and destroy (in the process of that industrial-level scanning), I was reminded of another AI-related article I hadn’t referenced yet.
In “The AI-Powered Web Is Eating Itself” from Noema this past January, writer Hamilton Mann explores how the AI arms race is destroying the some of the foundational aspects of the internet. More specifically, he points out how AI-powered search, now endemic to just about all the search engines out there, means a growing number users are finding more answers in the AI summary on the search page and never clicking on the links.
Not only does this seem like another example where AI search can start feeding on an ever-increasing amount of AI slop thereby increasing its breakdown of information into virtual grey goo, but it also cuts the information creators and curators out of the loop. Yes, I know we’re far from the early Internet where AOL was a walled garden anomoly. Now the tech titans prefer the defacto walled gardens of apps and the control they provide. But eroding and potentially destroying the free flow of information is, as Mann points out, going to be detrimental to both machine and human learning.
I’m still contemplating his proposal of “Artificial Integrity” and if that would work in the face of companies wanting to be data gatekeepers more and more.
And in the meantime, you should not forget about the cartoon villain monstrosity that is destroying rare and hard-to-find books that also could be taking away from the information commons. The AI companies have been doing this for a while now and knew enough to try and make sure this aspect of their business wasn’t widely known.
Hard to know what actions to take, but one step I am looking at is weaning myself off some of the standard search engines, which, frankly, have been getting gummed up by AI anyway. I’ve used DuckDuckGo and heard good things about Brave and StartPage for similarly anonymous searching. Many people have been very positive about Kagi, but that’s a (gasp) paid service. Suggestions are welcome in the comments.
What will the future of the series of tubes be?