Massive spike in "Illustrative Editorial" & "Similar Content" rejections?

I’m seeing a spike in "Illustrative Editorial" rejections

========

Are you sure that what you’re submitting meets Illustrative Editorial requirements?

Stock defines illustrative editorial as conceptual imagery designed to illustrate articles on current events and newsworthy topics. This type of content often features images of real brands and products — like signs on buildings, soda cans, computers, and cars — to convey a story. 

Illustrative editorial content can’t be used in commercial projects which significantly diminishes its appeal to customers. 

For illustrative editorial, Stock won’t accept:

• Images that feature recognizable people 
• Images of restricted events such as conventions and sports games
• Images that feature tight crops of copyrighted or trademarked material, such as stamps, fine art, or other content that may violate privacy rights
• Digitally created or manipulated versions of trademarked logos or other brand content — other than social media icons

https://helpx.adobe.com/stock/contributor/help/illustrative-editorial-content.html

In the past ~1.5 years, Adobe has definitely slowed down the growth of the database by various means, including capping the weekly submission for each Contributor, performing Audit Removals, rejecting many more for “similars”, and raising the bar on quality. Many of the rejections seem arbitrary and justified. Unfortunately, there’s nothing we can do as Contributors to overcome these hurdles. I have slowed down my submission rate dramatically, as the amount of work needed to capture, edit, upload, and add titles and keywords isn’t worth it if a high proportion of submissions are going to be rejected.

Regarding Buyers generating their own AI images, a certain amount of skill, experience and investment in software subscriptions is required to generate high quality, useful AI images. Many Buyers are under deadline pressure to complete a design and can probably find an image among the 400+ million in the database faster than they can design and edit their own AI image.

why would a buyer pay for an AI image that they can generate themselves in 30 seconds?

===========

Machines don’t create great content, people do.

• 95% of AI diffusions contain mistakes. 20-40% are unusable.
• The cost to generate a series of high resolution assets mounts up quickly. 
• It takes a trained eye to spot AI mistakes and above average editing skills to correct them in Photoshop, etc...
• Then the project must be upscaled and corrected again to meet Stock’s commercial quality standards.
• When all is said & done, that “30 seconds” has morphed into hours or days and various fees for multiple diffusion cycles and upscaling.

That’s beyond the pay grade of most average customers. AI content creation for Stock isn’t for everyone.

Stock does not and, never will, accept AI for Editorial use only assets.  

It’s true that Stock has higher acceptance standards than other services, and competition for shelf space is much steeper now than it was previously. 

The reasons for refusal are described in your Contributor User Guide. I suggest you re-read it as some things have changed.

https://helpx.adobe.com/stock/contributor/help/how-to-submit-distinct-content.html

https://helpx.adobe.com/stock/contributor/help/updated-artist-name-guidelines.html

https://helpx.adobe.com/stock/contributor/help/ip-guidelines.html

https://helpx.adobe.com/stock/contributor/help/reasons-for-content-rejection.html

https://helpx.adobe.com/stock/contributor/help/quality-and-technical-issues.html

For better feedback, we’d need to see 1-2 rejected assets & the reasons for refusal. 

Often, a 2nd or 3rd pair of eyes can spot things that you may have overlooked. 

but why would a buyer pay for an AI image that they can generate themselves in 30 seconds?

Some AI results are very impressive and unique, and licensing an asset from a stock platform offers better commercial security. If you generate an asset on your own, it is less likely to be fully cleared for commercial use.

Still, an asset approved on a stock marketplace does not guarantee it is free from issues. Even with a clean prompt that avoids real references, AI can still unintentionally violate image rights. Recently, I saw an AI-generated video of Jesus on a stock platform where he had the exact face of Keanu Reeves. It likely went unnoticed by both the contributor and the reviewers, but it is a good example of the commercial risks AI assets can carry.

“but why would a buyer pay for an AI image that they can generate themselves in 30 seconds?”

I would suspect that many are, but in my three years of creating AI assets, I could count on one hand the number of results that didn’t need editing whatsoever other than upscaling, while the rest required a minimum of 15 minutes and quite a few that required two to even 8 or more. It’s going to depend on the degree of quality a buyer is willing to accept and if the amount of followup time is worth the effort.

But yes, similar content and quality rejections are climbing for both photographers, illustrators and AI designers. I don’t know where the cut-off point is, but if there are somewhere in the area of over 2 to 3 million assets of a given subject, it’s going to get a similar content rejection unless the subject is presented in a unique way, with very specific and well-curated, carefully ordered keywords.