Administered by:
Talking with a few people about the apparent realism of generative AI outputs, I refer back to how Arthur Conan Doyle was taken in by the Cottingley fairy photographs, which to the 21st Century eye are obvious fakes
In a few years, will people be able to intuitively spot artificially generated images and text, even if not always able to articulate why?
@mattedgar Obviously I'm taking the piss there, but my point is that this stuff is invisible, until it isn't. I think there's a legit fascinating and very Gibsonian thing going here where humans are sort of flipping in and out of being able to tell as models iterate through various phases of sophistication. We're the dogs from the Terminator universe. No, now we're oblivious again. Wait, look at the fingers!
@mattedgar I had exactly the experience you mentioned recently where I saw a headline with a head shot photo of someone, and I was like, wait, that looks artificially-generated. Clicked through and the whole story was about a "person" who didn't actually exist. So there's definitely something there (I guess one can get all Gladwellian about your 10K hours or whatever). But will we be still be able to tell in future? Technological progress used to be slower (measured in years). I think.