The photography giant Getty Images has created an artificial intelligence tool that, based on texts written by its clients, will generate images inspired by the company’s vast content library.
The company assures that thanks to this tool and based on its own content, it will be able to create images with artificial intelligence without being affected by the still unresolved controversy over whether AI companies should pay copyrights to the authors of the works of the images. that they learn.
Getty, which owns the rights to millions of photographs, had even previously sued Stability AI, the company that popularized the Stable Diffusion image generator, for using its images without permission.
Getty’s new product, developed with chipmaker Nvidia Inc., will be based on Getty’s own content and will strive to sidestep thorny legal issues, in part by limiting the images that will feed the generator.
The new tool will take advantage of Getty’s creative image bank, but not its collection of news photos, as part of an effort to prevent the creation of deepfakes, says CEO Craig Peters.
The new image generator will not allow users to incorporate third-party copyrighted material or content they do not own, so there is no way to create something like the famous viral image of Pope Francis in a Balenciaga puffer coat, Peters assures.
To serve businesses looking to create ads and other types of content, Getty will allow customers to add their own material. Content generated through the product, which will create images based on text messages, will not be added to the Getty-owned content library.
AI-generated images will receive Getty’s usual license to use the content, as well as indemnification against lawsuits. The company also says it plans to compensate the artists and collaborators whose work it used to train the AI model.
The growing popularity of text-to-image AI, such as OpenAI’s Dall-E, Stable Diffusion, and the Midjourney service, has raised questions about whether such tools profit from the work of artists, photographers, and designers without obtaining their permission or compensating them.
Uncertainty around the technology means that companies that want to use artificial intelligence software to create new images for uses such as advertising campaigns or social media posts fear exposing themselves to legal risks and fines, says Peters, who noted that Getty clients have repeatedly raised this concern.
“There are real risks here,” he says. “Customers want to use generative AI, but they don’t want to be at risk of even knowing if they own that content.”
Earlier this month, Microsoft said it will defend buyers of its artificial intelligence products from copyright infringement lawsuits and pay any fines or settlements that may result.
Getty’s new service shows that AI companies that say they can’t develop the technology while respecting intellectual property rights aren’t telling the truth, according to Peters. “This fundamentally undermines one of the arguments of those who expose these generative models without compensation,” she said.