Getty Images, a leading provider of stock photos and videos, has launched a new generative AI service for its iStock customers. This AI-powered tool allows users to create unique visuals based on their own descriptions or prompts. However, Getty has taken steps to address concerns about AI models trained on copyrighted material generating infringing outputs.
This innovative service, powered in part by technology from Nvidia, is designed to prevent the generation of known copyrighted elements and is available in 75 languages. It can both modify existing images and generate new ones, and can be integrated with existing apps and plug-ins via an API. The cost for this service is $15 per 100 generated images.
The new service, called “Generative AI,” uses technology to prevent the creation of images that are substantially similar to existing copyrighted works. This is achieved by filtering out copyrighted material from the training data used to develop the AI model. Additionally, the service incorporates safeguards to prevent users from generating outputs that could be infringing.
Getty’s move into generative AI comes at a time when concerns about copyright infringement in the AI space are growing. Several companies have been sued for copyright infringement related to their AI-powered tools. For example, in 2023, a photographer sued Stability AI, the creator of the popular AI image generator Stable Diffusion, for copyright infringement.
Getty’s Generative AI service is a significant step forward for the company, and it could help to address concerns about AI-generated content infringing on copyrights. By using technology to prevent the creation of infringing outputs, Getty is ensuring that its customers can use its new service with confidence.