Adobe Wins Legal Battle Against Diversity Photos Founder Over AI Training Data
Key Takeaways
- ▸Adobe fed Diversity Photos' entire library into its Firefly and Adobe Sensei AI models without obtaining separate licensing agreements or compensation beyond a rejected low offer
- ▸Carter's legal challenge was unsuccessful; Adobe's interpretation of the 2018 agreement prevails, allowing it to use the archive for AI training
- ▸Adobe publicly marketed Firefly as respecting creator rights with opt-out options, but Carter claims creators in his partnership were never asked permission, offered opt-outs, or compensated
Summary
Gerald Carter, founder of Diversity Photos, has publicly disclosed his failed legal challenge against Adobe over the company's use of his image library to train its Firefly AI model. Carter claims Adobe used his entire photo archive—containing rare, diverse imagery built over years—without explicit permission, compensation, or opt-out options, despite publicly claiming it respected creator rights. Adobe successfully used legal resources to thwart Carter's challenge and is now attempting to use the arbitration ruling as precedent for future disputes. The dispute stems from disagreement over the interpretation of a 2018 Stock Contributor Agreement between the two parties, with Carter arguing it only covered distribution rights, not AI training rights.
- The dispute highlights ongoing tensions between AI companies' training practices and content creators' intellectual property rights, particularly regarding rare or specialized datasets
Editorial Opinion
This case underscores a troubling pattern where AI companies leverage broad contractual language to justify uses that extend far beyond original intent, while simultaneously marketing themselves as creator-friendly. Adobe's legal victory may prove pyrrhic if it sets precedent that erodes trust with content partners and creators, particularly those in underrepresented communities whose work carries specific value and mission-driven constraints.
