As someone who is largely around the art community admiring and sharing thier work, the fact that I could confuse AI Generated Images and thusly falsely share or save them has been such a huge anxiety of mine every since 2022
One easy way to check is the look for JPEG artifacts that doesn’t make any sense. A lot of the systems were trained with images stored as JPEGs, so the output will have absurd amounts of JPEG artifacting that will show up in ways that make no sense for something that actually went through JPEG compression, such as having multiple grids of artifacts that don’t line up or of wildly different scales.
I’m really bad at noticing small details. Luckily 99% of AI artists use the same art style (with more or less Pixar influence for humans) so I can still spot AI imagery from a mile away
I’ve had moments where they admitted to Generating the Images in thier Bio, yet even with that knowledge I could not tell. I reccon this is much more of an issue in the Anime Artist scene where there are more varied Art styles to steal and replicate…
As someone who is largely around the art community admiring and sharing thier work, the fact that I could confuse AI Generated Images and thusly falsely share or save them has been such a huge anxiety of mine every since 2022
One easy way to check is the look for JPEG artifacts that doesn’t make any sense. A lot of the systems were trained with images stored as JPEGs, so the output will have absurd amounts of JPEG artifacting that will show up in ways that make no sense for something that actually went through JPEG compression, such as having multiple grids of artifacts that don’t line up or of wildly different scales.
Confuse AI with what?
How do you “falsely share or save”? I think every time I shared or saved something it worked. I could be wrong.
(I’m in art too. Or was. I’m sorta changing these days.)
I’m really bad at noticing small details. Luckily 99% of AI artists use the same art style (with more or less Pixar influence for humans) so I can still spot AI imagery from a mile away
And the face is always one of these.
All of these faces make physical sense, while AI art often doesn’t.
Or you only notice the obvious ones and are oblivious to all the ones you have not recognized
I’ve had moments where they admitted to Generating the Images in thier Bio, yet even with that knowledge I could not tell. I reccon this is much more of an issue in the Anime Artist scene where there are more varied Art styles to steal and replicate…