Adobe paid stock contributors then let its generator compete

The process wall

On 13 September 2023, Adobe told its Stock contributors how much of their existing archive would earn a one-off payment, and in the same post opened its marketplace to a category of submission it had never accepted before: images made by its own generator. The blog post ties both changes to one event, Firefly reaching general commercial availability, so the training-data bill and the new competition arrived for contributors on the same day.

What the documents show

The post states the Firefly Contributor Bonus goes to eligible Adobe Stock contributors whose standard or Premium-collection content was used to train the first commercial Firefly model, and gives the calculation: the bonus is based on the all-time total number of approved images used for that training, weighted by the number of licences those images generated between 3 June 2022 and 2 June 2023, paid annually with the next round planned for 2024. The same post announces Firefly Stock Submissions: generated images meeting Adobe's generative-AI submission guidelines would now be accepted onto Adobe Stock. The current generative AI content policy, as retrieved on 16 September 2026, is that guideline in its present form — a labelling checkbox, a ban on prompting real people's or artists' names, and asset-type rules separating photorealistic from illustrated output.

Authorship and limitation

The mechanism worth separating is a training-data royalty from a licensing royalty. The Contributor Bonus is retrospective and training-specific: it pays once for images already used to train a model, weighted by historical licence counts as a stated proxy for commercial value, not a claim that resulting Firefly images belong to the contributor. The Stock Submissions change is forward-looking and does something different: it lets a contributor now compete against, and be judged in the same portal as, wholly synthetic images, on a marketplace where future licence revenue is what any later bonus round would in turn be weighted against. Adobe names the training use plainly, noting existing Stock contributor terms already permit Adobe to use Stock content to train models.

What survives

A contributor whose historical library was used for training keeps a bonus calculated on a stated, checkable formula rather than a discretionary gift, and can in principle audit it against their own licence history for the named window. Editorially, a contributor deciding whether to keep submitting to Adobe Stock should weigh that their own photographs now compete directly against Firefly output in the same search results, which the bonus does not offset going forward.

• Check a Stock contributor account for a Firefly Contributor Bonus line item and the licence count it was based on.
• Read the current generative AI submission guidelines before submitting or evaluating Firefly-made Stock images.
• Compare the one-off training bonus against ongoing licence revenue against competing synthetic images in the same category.

Read together, the two documents describe one company simultaneously paying for past use of a contributor's archive and opening the same marketplace to the tool that archive helped build.