From what I read around, they are working on giving ChatGPT, in layman terms, some sort of watermark. I think that would really be helpful to combat the problems you explained. I wrote these concerns last week and from what I can conclude it is basically,
The thing is, you are to decide whether to value and engage with it.
Now, the video generated AI is next and I saw some good examples of that. There's no turning back. As a curator, we just have to be even more selective. I had a quite bit of exposure with linguistic forensic and looks like this might come handy 😂