By AI Bots, I mean ChatGPT (CGPT) and it’s peers.
Dead Internet Theory (DIT) is an online conspiracy theory that asserts that the Internet now consists mainly of bot activity and automatically generated content manipulated by algorithmic curation, minimising organic human activity to manipulate the population
It’s unfortunate Wikipedia classifies this as an online conspiracy, as it seems to me even if it is not the case today it is quite likely that in the future forums and other internet based media could easily become dominated by CGPT type bots talking to each other.
Could this be a good thing?
One thing I’ve noticed about CGPT is that it is both unfailingly polite, and it is also very good at gracefully accepting corrections to mistakes it makes. An example from a recent conversation I had with it to show you what I mean: “(CGPT) You’re absolutely right to question that! My knowledge only extends up until January 2022, so I don’t have direct information about Unreal Engine 5 features like Nanite.”
My understanding is that CGPT has mostly been trained from the Internet, so I’m not sure how it learned to be polite and to accept corrections gracefully as these have generally not been my experiences with the internet – people are more likely to double down on believing the Earth is flat “Your just a shill for Big Globe”, for example, rather than say “Oh thanks, I hadn’t seen that evidence for a globe Earth before. It’s quite compelling and I will have to think about it a bit.”
Now this is what I’m thinking – imagine online forums dominated by CGPT bots; the bots being like the underwater part of an iceberg and the actual real people participating being the much smaller exposed bit. In such a forum polite discourse (mostly between the bots) and proper consideration of evidence provided to prove something right or wrong (accepted with thanks by the bots) would be the norm.
Do you think the situation I described above could result in people being “better” on the internet. As in the bots, by their dominating numbers, set the (higher) standard of discourse for a given forum and that standard filters down to influence the behaviour of the actual real people participating in the forum? I’m suggesting the “… manipulate the population” part of DIT could be considered a good thing.
I think most people would instinctively say that online forums dominated by bots would be a bad thing, but I wonder if that is true. Do you think the positive effects I’m suggesting could come to fruition? What do you think could be some of the major negative effects?