Sell photos to AI: who takes stills and how they pay

Which buyers take photos for AI training?

Shutterstock's help page for contributors says its datasets currently hold visual content including images (photos, illustrations, vectors), videos, 3D models and music.

Source: Shutterstock, data licensing and the contributor fund

Wirestock tells creators that the first step is uploading photos, videos and illustrations to it, just as they would for traditional stock licensing.

Source: Wirestock, earning from AI training datasets

Adobe Stock's Firefly FAQ for contributors puts the question of whether Adobe compensates contributors when Stock assets train Adobe Firefly, and names a Firefly Contributor bonus program. The page we read gives no rate for it.

Source: Adobe Firefly for Contributors FAQ

How are photos paid for when an AI company licenses them?

Shutterstock pays from a pool: earnings from datasets go into a collective fund that it distributes periodically, and its contributors earn a 20% average corporate royalty rate of what it receives for data licenses. That is the only percentage for stills on the board.

Source: Shutterstock, data licensing and the contributor fund

Wirestock says that once your data is selected and used by AI companies you earn royalties or lump-sum payments. Neither of its two pages we read puts a figure on either.

Source: Wirestock, earning from AI training datasets

What about the other buyers on the board?

On the pages read up to 2026-10-08, what Claru, Perle, GoPro, Defined.ai and Troveo ask for is video: first-person capture, action-camera clips and owned film footage. Protege names media among its fields without saying whether stills are part of it, so for a photograph the stock libraries above are the place to start.

Source: What AI buyers pay for footage, the board

If people can be recognised in your photos, the consent questions on the page about footage with people in it apply to a still in the same way.