[shagie]

This is going to be the HeLa problem for future AI. ( https://en.wikipedia.org/wiki/HeLa#Contamination )

If we contaminate the content that we are training future AIs off of with ChatGPT, then they'll have problems with the value of that information in the future.

I'll also note that if one attempts to purposely contaminate information with ChatGPT to do that "make it worthless to future AI", it also makes it and the tangental or adjacent information worthless to humans too.