Getty Images has signed a multi-year deal with OpenAI that will bring its licensed photo library into ChatGPT and OpenAI’s search results. The agreement, reported by Engadget, gives OpenAI access to Getty’s visual content, though details about whether those images can be used for AI training have not been disclosed.
“High-quality, licensed visual content makes AI-powered search and discovery more useful and more trustworthy,” Getty CEO Craig Peters said in a statement. “This partnership with OpenAI reflects a shared recognition of that, and together we will deliver richer visual experiences to ChatGPT users.”
The deal is notable because Getty spent years taking an aggressive stance against AI companies. That position has clearly softened, and the shift tells you a lot about where the industry is heading.
For a long time, Getty was one of the loudest voices pushing back against AI’s use of copyrighted images. In September 2022, it banned all AI-generated art from its library. A few months later, it filed a lawsuit against Stability AI, alleging the company had used millions of Getty images without permission to train its image-generation model. That case was eventually rejected late last year.
Then came the pivot. In 2023, Getty launched its own generative AI image tool, trained on its own library and built on NVIDIA’s Edify AI model. Images produced by the tool came with royalty-free licenses, which helped Getty sidestep the copyright issues that had tangled up its competitors.
By October 2025, Getty had signed a content deal with Perplexity AI, allowing that company’s search tools to pull from Getty’s library. That agreement included a commitment from Perplexity to display proper image credits and links back to the source, a detail that mattered given Perplexity had faced its own allegations around unlicensed use of copyrighted content.
The OpenAI deal follows that same path. What’s still unclear is whether OpenAI will be allowed to use Getty’s images for training its models. Getty’s deal with Perplexity explicitly ruled that out, but no such clarification has been offered here. That gap is worth watching, especially as copyright litigation around AI training data continues to play out across the industry.
For OpenAI, adding licensed visual content to ChatGPT is a practical move. Search tools that return images alongside text answers are more useful, and sourcing those images from a legitimate, well-known library reduces the legal exposure that comes with scraping the web. For Getty, the deals with Perplexity and now OpenAI suggest the company has decided that working with AI platforms, on its own terms, is a better strategy than fighting them in court.
Whether that calculation holds depends on what the fine print actually says, and Getty has not shared much of it.
