I know AI is improving all the time and, despite how we feel about it, it's here to stay.

The main sites have made their choices to accept it or not, and I'm sure most of us are working within the guidelines given.

However, this month on Shutterstock has gotten off to a really bad start, so I did some searches to see what has happened to the images that were selling daily, but have now suddenly stopped.

I looked at some broad subjects, where my images were well placed, and there is now so much approved AI, that real photos are being pushed further and further back. I'm not referring to Shutterstock's AI tool, but to people who are trying to pass off AI as real photos.

I have a lot of wildlife shots, so the first searches I checked, (restricting to photos and any time) were as follows:

'Koala'  - nearly 25k images. The number 1 spot is an Offset image and images 2 - 6 are AI, with several more on the first page.

'Kangaroo' - over 71k images. The top 10 has 3 AI images, and the 2nd image is a very rare 6 legged kangaroo.

'Lion' - over 600k images. At least 9 of the first 16 are AI.

'Jellyfish' - over 86k images. The top 10 are 2 from Offset and 8 AI.

'Polar Bear' - 44k images. At least 5 of the first 13 are AI, with AI images taking the first two spots.

'Dolphin' - 90k images. Of the top 30 images, no less than 15 are AI.

For every search I make, I'm finding similar results, and I've reported several portfolios to Shutterstock. A few have been removed, but many remain.

I know that someone started a thread here with stolen content, which makes sense, as we all want to check such portfolios for our own images, but I'm hoping that by listing some of the AI portfolios here, Shutterstock might take notice and take action quickly.

Here are a few that I've found today:

https://www.shutterstock.com/g/munawaekhan
https://www.shutterstock.com/g/almogphoto
https://www.shutterstock.com/g/Amarphotocreater
https://www.shutterstock.com/g/Tahseenamjad
https://www.shutterstock.com/g/Toufiqrahaman
https://www.shutterstock.com/g/Behramkhan03
https://www.shutterstock.com/g/rinku+makwana
https://www.shutterstock.com/g/ContentCreater1
https://www.shutterstock.com/g/IrfY
https://www.shutterstock.com/g/MuhammadAsif6


Please feel free to add any you find to this thread, or report directly, as these images are forcing us down in the searches, and it's not restricted to wildlife, but to almost all subjects, from food to interiors to landscapes to Christmas.

How 6 legged kangaroos are getting through a review process, is beyond me, so I can only assume that perhaps AI is checking on AI? Maybe if we start to call it out here, someone at Shutterstock will take action?

Yes, some blatant AI imagery in those ports.

I know SS have opened up the flood gates in their acceptance, but how on earth are the SS reviewers allowing them to pass?

I also mentioned this in another thread but if they are using our content for AI training, doesn't using AI images end up degrading the model? (unless companies like Nvidia have their own AI detection tools).

Also, are there any free online tools to check if an image was AI generated?

SS partnered with OpenAI to provide AI image synthesis services using the DALL-E API. Anything that comes from SS is from OpenAI. As for training from AI to teach AI, that's 100% wrong and shouldn't be done as it defeats the natural progression of "training".

Passing these horrid images, and if they aren't marked as AI, then the whole system is broken on SS. AI reviewing AI is just as bad.

In the past they turned a blind eye to people using multiple keywords in the title to game the search, I doubt they are any better or care more now.

But customers licensing images are going to complain pretty loudly when they realise that the kangaroo in the 'photo' has two extra legs. 

I don't know if there are any free tools but, as Shutterstock strip the metadata, then it's probably not going to allow us to check submitted content, but they are a tech company, and should certainly be able to spot any markers in uploaded files. There reviewers should also be able to spot the obvious here, so they either don't care or don't exist, and the checks are Ai?

The ones I added above are just the tip of the iceberg.

Yeah I've seen plenty of them including video.

Kind of makes you wonder what's coming next from SS, maybe another 'exciting news' email.

I've just browsed through lots of photos and videos that used AI in some shape or form, I knew AI was getting through the cracks but had no idea it was this bad.

So think it's confirmed, SS are officially unofficially accepting AI content, there's no denying it now.

Get ready for an official announcement or maybe a soft launch where they don't tell anyone, but quietly change their terms and conditions.

So their AI reviewer failed to spot the obvious AI slop. It's still a win-win for SS. Once these ports get reported and banned they get to keep all the earnings