Shutterstock and OpenAI are extending their collaboration for another six years, during which time the AI startup will be able to use the vast collection of photos, videos, music, and metadata from Shutterstock to train its models.
The collaboration between the stock photo website and OpenAI originally started in 2021. At that point, Shutterstock began allowing OpenAI to utilize its photos to train DALL-E, a text-to-image model. Sam Altman, CEO of OpenAI, called this agreement “critical” to the model’s training. Last year, Shutterstock introduced a “Contributor Fund” to pay when an artist’s work was used to train an OpenAI model.
At same time, Shutterstock also integrated OpenAI’s image generator directly into its website and forbade the selling of AI-generated photos produced using anything other than the DALL-E tool that was already embedded into its platform. However, Shutterstock is now increasing this integration and claims that doing so will allow users to “edit and transform any image in the entire Shutterstock library.” Giphy, the platform for creating GIFs that Shutterstock just purchased from Meta, will also get AI features from Shutterstock.
Brad Lightcap, the chief operating officer of OpenAI, says in a statement, “We’re happy to be able to license Shutterstock’s high-quality content library. In addition to improving the capabilities of our picture models, this expanded partnership gives brands, digital media, and marketing firms the keys to unlock revolutionary possibilities for ideation and content production.”
Shutterstock, in contrast to other image-sharing websites like Getty Images, is completely embracing AI and any potential consequences. Getty Images has responded to artists’ concerns about their work being scrapped to train AI models by fully prohibiting AI-generated content from its platform. Additionally, Getty Images filed a lawsuit against Stability AI, claiming that it had “illegally copied and processed millions of images protected by copyright” on its website.