Reporting AI Images That Haven't Been Labeled As Such

[iedreams]

Most of the images in this contributer's account are labeled as being AI generated. 

But the images in this series (linked to below) are not, even though they most likely are AI generated as well. 

https://stock.adobe.com/search?filters%&serie_id=572848402

[Jill_C]

Adobe no longer requires Contributors to label their Generative AI submissions, neither in the title nor the keywords; however they must select the Generative AI checkbox when uploading. Apparently, the filter does a good enough job of detecting if the asset is Generative AI, allowing the Buyer to either include or exclude them in their search.

[Justin2933350612u3]

I also found yet another contributor who is not labelling images as generative Ai even though they obviously are. Marketing department picked an image I had to work with that is obviously generative Ai (dead giveaways - Absolute poor path work, artifacting, details fall apart upon closer inspection, etc. Just like other generative Ai vectors.) Looking at the contributors portfolio some are labeled as Ai but not all. Portfolio is also a dead giveaway with all the similar iterations. 

Image: https://stock.adobe.com/images/beautiful-autumn-mountain-landscape-amazing-landscape-of-snowy-mounta...

Contributor's portfolio https://stock.adobe.com/contributor/211732338/rousan?load_type=author&prev_url=detail

It would be great if there was some form of curating for new images being added to Adobe Stock and a simpler way to report images/contributor since this appears to be a pretty common problem and is only getting worse. Trusting the contributors is not working. Besides there cannot be a large market of people specifically looking for generative Ai when it can be easily done and for little to no cost done by the person searching themselves. 

[Jill_C]

Based on the asset number, this image appears to be a bit older - at least 200 million images have been added to the database since this one. I think Adobe Stock has more automation to detect improperly labelled AI assets now. Perhaps @Contributor1 can send this one back to the Moderation Team for another look.

I think that your premise that there is not a large market of people looking for Generative AI images is flawed. There are  now more than 100 million AI images in the Adobe database, perhaps more if you include those that aren't labeled as such. Adobe wouldn't invest significant resources in acquiring and hosting these assets if they weren't experiencing significant AI sales. As many in the Gen AI creation business have learned, creating a Gen AI image is easy, but creating a high-quality saleable image takes work and patience and a knowledge of various tools. That's why designers come to Adobe to license such assets rather than trying to do it themselves.

[Contributor1]

Hello @Justin2933350612u3, I see the issue you're referencing I've sent these details to be reviewed internally. Thank you for the report and sorry for the problem with the vector

[jessel61795183]

I'm designing an aquarium. I need realistic, biologically accurate images. If I accidentally put an AI generated image into a presentation, this reflects poorly on me. And ultimately reflects poorly on adobe. Sure there is a place for AI generated images/videos. But some of us really need to be able to filter those out. Here is an example of a video not filtered by "exclude generative AI" that is definitely AI:

https://stock.adobe.com/video/sea-anglerfish-not-in-a-predatory-stance-but-floating-serenely-in-the-...

This kind of thing just wastes my time and makes it harder to find what I actually need. Hopefully Adobe will actually try to fix this problem.

[Jill_C]

Perhaps @Contributor1 can have the moderation team take a look at this one.

[Julian331072003sgd]

How is this the standard procedure to mark non-labeled AI generated images? I stumble upon dozens of these when trying to find stock footage. Should I open a new discussion in the community board everytime? What kind of paperwork deterrent is that supposed to be? 

My coworker unknowingly just used an AI-generated picture for a customers homepage. We are by law obliged to mark AI generated pictures - but how could we if we aren't aware? And how can we prevent that from happening if there is no simple process to flag those pictures?