Skip to main content
Participant
July 3, 2023
Open for Voting

Add report option for incorrectly tagged AI-generated images

  • July 3, 2023
  • 4 replies
  • 356 views

I've seen several other posts in other threads saying this but I just wanted to add my voice to the chorus and say that as a user of stock photography I'm sick of AI-generated images showing up in my searches even when I've selected to exclude them in the filters. There are certain users who keep showing up in my searches as doing this (when you go on their libraries it's wall-to-wall AI stuff, so I have to imagine it's an intentional thing). 

 

From what I've read this isn't against the rules currently but I just wanted to say I think it should be, and I'd like to be able to report or flag these posts. I know I can't do much as an artist against the tide of AI art but the least we can do is keep it separated from the more thoughtfully created stuff.

 

Thanks!

4 replies

daniellei4510
Community Expert
Community Expert
August 14, 2024

Adobe is canceling AI accounts left and right. We see it here everyday, with someone asking why their account (or even accounts) was/were deactivated or banned. There's just so much crap, and I suspect Adobe is having trouble keeping up. Others have suggested a report button, but I think Adobe is concerned it would result in people making false reports. But just a guess.

Adobe Community Expert | If you aren't submitting your assets in sRGB, you probably didn't read the rules.
Participant
August 14, 2024

Adobe is profiting, they do not care about the user just profits. 

Abambo
Community Expert
Community Expert
July 4, 2023

Adobe claims that they detect AI thanks to AI, but that does not work in all cases. Having some kind of report button could be helpful, but how to protect it from abuse?

ABAMBO | Hard- and Software Engineer | Photographer
fabrice+ducks
Known Participant
July 31, 2026

It seems to me that it would be a better workflow for Adobe to detect AI-generated files based on technical specifications rather than graphical details.

For example a script could open a .EPS or .AI file and inspect the layers panel. If every single element is tucked inside the same folder, chances are this illustration is AI generated. 
If the folders have names, structure, chances are it isn’t.

There are probably other technical determinants (for bitmap files for example), but the mass identification of vector files shouldn’t be too difficult.

There is another possibility which is that users could flag an image for suspicion of AI generated nature. In this case a user could filter out images with 10 or less flaggings, or 50 or less flagging, and above a certain number (if 200 people flag it?) the image would be removed from search engines, with a possible appeal process for the creator.

Just a few ideas.