Getty Images bans the upload and sale of AI-generated images using tools like DALL-E, Midjourney, and Stable Diffusion, citing concerns over legal challenges
Getty Images has banned the upload and sale of illustrations generated using AI art tools like DALL-E, Midjourney, and Stable Diffusion.
Context & Ripple Effects
This ban is the opening move in what became Getty's full confrontation with generative AI: within months it moved from refusing AI images on its own platform to attacking the models themselves, first with planned UK legal action over copyright, then a US suit alleging Stability AI trained Stable Diffusion on more than 12 million of its images without permission or compensation (the US lawsuit) — and eventually a bid to halt UK sales outright.
The strategic contrast was immediate: Shutterstock signed with OpenAI to sell DALL-E 2 output with a contributor fund (Shutterstock's OpenAI deal), and Adobe took the middle path of selling artist-made generative work so long as it is labeled (Adobe's labeling policy). Getty bet its catalog's value on provenance; its rivals bet on distribution.
First-order effects
- Contributors lose a sales channel overnight — any Midjourney, DALL-E, or Stable Diffusion output is now unsellable on Getty, pushing that supply toward Shutterstock and Adobe Stock.
- Buyers who need defensible licensing for commercial use get a cleaner corpus at Getty, but pay for it with a library that excludes the fastest-growing category of new illustration.
Second-order effects
- Shutterstock and Adobe convert Getty's refusal into a differentiation play: labeled generative inventory plus creator-compensation schemes becomes their pitch against Getty's all-human catalog.
- The ban hardens Getty's litigation posture — once the platform refuses to host these tools' output, suing the model makers over training data (as it did against Stability AI) reads as one coherent strategy rather than an isolated dispute.
Third-order effects
- If the pattern holds, the stock-image market splits structurally into two camps — platforms that govern and label generative supply versus platforms that exclude it and litigate against model builders — with provenance becoming the core product feature.
- Model makers like Stability AI face a business-model squeeze: open training datasets invite platform-level lawsuits, pushing the industry toward licensed or compensated training data as a precondition for commercial text-to-image tools.
The trend: Stock imagery is splitting between platforms that commercialize labeled generative content and platforms like Getty that exclude it while pursuing the model makers over training data.