AI is a scam
It seems no matter where you go, There's this talk about that "AI" taking over. And it has been going for a while right now, but it ramped recently up becuase you have stuff like midjourney or chat-gpt. I'm kinda tired of it and think the way people talk about it is bullshit.
There are the so called "dangers" of AI. And I find them laughable. You have all of these stories like robots becoming super intelligent or killing humanity or whatever else coming up. I think the simple response of "bullshit" would suffice to put these things to sleep. But if I had to take them more seriously. All AI does is pattern matching. The big models are just so stuffed with data they can mash things together and appear coherent (if you don't rock the boat too much). And pattern matching can get you far, but it's just a subset of all ways you can think. I would even say it's borderline not thinking, but more closer to just reacting to stuff. As for the comment that this is just what humans do too. well, people don't need to be fed billions pages of text before they appear coherent. You don't even have the time to do it if you take into account how long it takes to read a page. People are also able (or at least used to) to create wholy different worlds if they want to. There were times when planes didn't fly, when painting styles like cubism would seem like from a different star system, and when the idea of reading and writing didn't exist. These were qualitative changes and not a simple mash ups inside a known space.
But going past sci-fi cliches, but there are real dangers. Just not this. What real dangers (I imagine) are is the following:
Lack of coherence
A good chunk of what improves information is being able to check where it's coming from. It's why sources in books are a good thing. If you read something that's 'beyond belief', you can check it for yourself where it's coming from. That way you actually have a coherent narrative. You can't bullshit your way through things, at least not easily. AI goes away with that, as you are sold the illusion it's the bot "answering" what you put in instead of mashing stuff that was already written together. And what way could you have to verify whatever is mashed together has some semblence to reality?
Of course, you could have these systems provide where they are getting their stuff from. But that would kinda kill the illusion of AI, no? It would become what I think it is, a text masher.
An other thing to consider is whenever you aren't killing off what gave you data in the first place. If you replace people with AI, you are essentially cutting of people from providing data to the system. Which begs the question, is the AI just going to keep doing whatever it's doing when you trained it for eternity? We never did things the same way. Building methods were different a century ago, so was fashion or even the way people speak.
You can't really have the whole world running on bots. One reason would be it's economically absurd. Having humans buy stuff you produced with bots makes sense. Having bots buy stuff from other bots makes you wonder why the bots are even running. So you will have to put people somewhere in the economic reality of the situation.
Of course you might say that whoever owns the AI will just live like a king while everyone else just will get poor or put on UBI. Which yeah, seems what currently is likely to happen, but that's not a stable social aragement. Unless you want some kind of revolution to take place in following decades.
But returning on the question of no new data being fed. Maybe AI will just lead to nostalgia suffocation. Which already is and was happening without it. Every day I read about some movie or game which I saw 20 years ago being remade. People actually dress the same today as they did 10 years ago too. Or at least very close to that. There used to be times when a decade skip felt like entering a different dimension. Think about music from 60s and 70s, and then thing what was playing 10 years ago compared to now.
If songs now didn't have release dates with them, you wouldn't be able to guess which year they came out. But when you think about what made it possible for people to experiment in the past, it's actually having a pretty comfortable living situation. You didn't have to worry about your job as much decades ago as you do now, and neither was there a whole overarching apparatus to keep you in line. So you could fuck off and do your thing.
Can you still do that in age where you are constantly monitored, filtered, and made poorer by machines?
Last one is how much bullshit is being created by these things. You now have, no matter how silly it is, people who are feeling insecure about their faces because of the AI filters to the point of tears. Which is kinda funny, but I don't think we need more depressed teenagers scrolling through their phone in this bizzare world wide game of looks.
I imagine it will get worse. There will be lot of people who will try to replace human relationships with these chat bots, and as for what's wrong with that. I like having people who surprise me and I can learn from, but this was much better answered over a century ago
Of course you can have these thing used for good too, but I feel to do it you kinda have to stop thinking in terms of AI and see these things as just data mashers. Which for many tech people would seem like an unholhy act