Getty Images and NVIDIA are deepening their artificial intelligence partnership with the launch of GenerativeAIbyiStock, a text-to-image platform specifically designed to create stock footage.

iStock’s GenerativeAI builds on Getty’s first foray into AI image generation, GenerativeAI for GettyImages. The difference is that the image platform of iStock, a subsidiary of Getty, provides assistance to individual users, while GettyImages is different and is more like a multi-user enterprise solution.

iStock’s GenerativeAI is trained using NVIDIA’s Picasso model, learning exclusively from Getty’s Creative Library and iStock’s Image Library. It was not trained on Getty's library of editorial images to prevent the generation of trademarks or known people.

"It makes the user's workflow more efficient and the photos obtained more accurate." Grant Farhall, chief product officer of Getty, said that iStock's GenerativeAI is mainly targeted at small and medium-sized enterprises that need to find stock photos.

"It allows users to be more efficient in their workflow and get the more precise photos they need, even things they can't do with a camera," Farhall said. He gives the example of someone looking for photos that illustrate climate change: They can use iStock's GenerativeAI to create a photo of penguins walking through a city street; instead of hiring a photographer to find a group of penguins, the AI ​​can create that photo for them.

Pricing is $14.99 for 100 tips, with each tip generating four images.

Another big difference between the GettyImagesAI platform and iStock's new service is legal compensation. Unlike GettyImages’ GenerativeAI, users cannot receive unlimited compensation. The iStock platform is capped at $10,000 per asset, which is the same license it offers for existing stock photos. As with Getty’s first generative AI platform, customers can participate in a revenue sharing program based on Getty’s traditional licensing revenue program.

The platform also offers new Inpainting and Outpainting features. Inpainting allows users to block a certain area of ​​the picture and then fill it in with people or objects based on text prompts. Outpainting expands photos according to different aspect ratios and fills these new areas.