Skip to main content
BlueShadow2049
Participant
February 4, 2026
Question

increase of Rejection rates

  • February 4, 2026
  • 3 replies
  • 123 views

When I first started submitting AI-generated content to Adobe Stock, acceptance rates were around 90%, even though overall AI quality at that time was far lower than today. In the early stages I was submitting only 4K AI videos and they were consistently accepted.
 

About a year later, my acceptance rate dropped to roughly 50, so I adapted and began submitting HD instead of 4K. Recently, however, things have worsened again, over the past month acceptance has fallen close to 10% for both AI images and video, even though I am still submitting HD and maintaining the same production standards.


What’s concerning is that during this entire period, AI model output quality has improved massively, and my own selection and QC standards have become much stricter, yet acceptance has continued to decline.


I’ve also significantly reduced my upload volume and now submit around 10 videos per day, all carefully curated and manually processed. Despite this, acceptance has not improved, and most recent rejections are labeled simply as “Quality.”


At this point I see no option but to stop submissions.

Based on what I’ve seen reported in forums, contributors who have only recently started submitting AI content are often mentioning much higher acceptance rates. From the outside, this makes it seem like the steep decline may be affecting long-term, higher-volume AI contributors more than newer ones.

Is anyone else seeing a similar long-term downward trend in AI acceptance rates?

    3 replies

    Nancy OShea
    Community Expert
    Community Expert
    February 5, 2026

    Have you tried submitting Stock rejections to other outlets?

     

    Nancy O'Shea— Product User & Community Expert
    Nancy OShea
    Community Expert
    Community Expert
    February 4, 2026

    Is anyone else seeing a similar long-term downward trend in AI acceptance rates?

    =============

    I think your analytics are mistaken. Adobe Stock continues to accept two to three times more AI generated content compared to other content types. 

    But customers buy only a small percentage of it. Some customers filter out AI from search results.

    Therein, lies the problem.  Over supply vs lower demand.  🤔

     

    To succeed at Stock, it’s not enough to be good. Your work has to be exceptional. The competition is steep, and it’s getting steeper every day.

     

    Maybe you need a change of pace. Instead of working on AI video, try something else. Find another niche that hasn’t been oversaturated with competing content.

     

    Good luck.

     

    Nancy O'Shea— Product User & Community Expert
    BlueShadow2049
    Participant
    February 5, 2026

    Thanks for the suggestions.

    In my case, sales are not the issue, my AI content performs well and I focus on exceptional-quality production, so limited AI demand hasn’t been a major factor for me.

    My concern is specifically about the sharp increase in rejection rates, which I described in detail in my original post.

    RALPH_L
    Community Expert
    Community Expert
    February 4, 2026

    I think you are wrong in your assumption that new contributors are given “breaks”. I do however, know that the review of AI generated assets has gotten tougher. HD over 4K may (I dont really know) increase overall quality, it does not prevent the rendering errors that we constantly see here in this forum. 

    BlueShadow2049
    Participant
    February 4, 2026

    Hi Ralph,

    Thank you for the clarification. That’s helpful.

    Since you mentioned rendering artefacts, could you share what types are most commonly triggering rejections in AI video right now? I already run heavy QC and traditional cleanup tools (e.g Topaz denoise/upscale), but that clearly isn’t sufficient anymore.

    Do you see more issues with things like temporal flicker, facial warping, texture crawling, lighting inconsistencies, or sharpening halos? And are newer “diffusion-based” models becoming necessary rather than classic post-processing alone?

    Any specific pointers would be greatly appreciated.

    RALPH_L
    Community Expert
    Community Expert
    February 4, 2026

    Sorry, I cannot help you with videos. My expertise is with photos.