The internet boasts a staggering repository of over 15 billion AI-generated images. Contributing to this vast collection of AI-forged media is Getty Images, a prominent supplier of stock imagery on a global scale.
This U.S.-based media powerhouse announced their intention to introduce generative AI capabilities into their creative content repertoire. This groundbreaking move signifies that users of Getty Images’ services will soon wield the power to manipulate and craft images with the aid of AI within the creative section. Notably, this tool utilizes NVIDIA’s Edify model, a part of the Picasso generative AI model library, trained exclusively from Getty Images’ creative library.
Getty Images maintains an extensive collection of editorial content primarily catering to newspapers and magazines, focusing on real-world figures and current events. It’s important to note that the generative AI tool will not be applicable to editorial content. As a result, users won’t be able to employ images of prominent figures like Joe Biden or Elon Musk in AI-generated visuals.
Generative AI by Getty Images operates in a fashion similar to Adobe’s Generative Fill in Photoshop, enabling users to generate content based on text prompts. Getty Images has released a video demonstration to illustrate the tool’s capabilities.
Craig Peters, CEO at Getty Images, expressed his enthusiasm about this launch. “We’re excited to launch a tool that harnesses the power of generative AI to address our customers’ commercial needs while respecting the intellectual property of creators,” said Craig Peters, CEO at Getty Images. “We’ve worked hard to develop a responsible tool that gives customers confidence in visuals produced by generative AI for commercial purposes.”
Licensing and copyright
Regarding licensing and copyright, Getty Images offers three types of content license agreements: royalty-free, rights-managed, and rights-ready. Users generating AI images on Getty will receive the standard royalty-free license, which includes various provisions such as representations and warranties, uncapped indemnification, and the right to perpetual, worldwide, nonexclusive use across all media.
Notably, content generated through the tool will not be integrated into existing Getty Images and iStock content libraries for third-party licensing. The company has also affirmed that contributors will be duly compensated if their content is included in the training set.
The generative AI explosion
The proliferation of generative AI has become increasingly evident across diverse industries. Notable instances include OpenAI’s implementation in ChatGPT and DALL-E, Amazon’s integration into Alexa, Adobe’s enhancement of Photoshop, and Apple’s utilization in their latest product releases.
The surge in generative AI models, including DALL-E, DALL-E 2, and DALLE-3, Midjourney, Stable Diffusion, and Adobe’s Firefly, has resulted in the creation of over 15 billion images. This figure surpasses the number of images available in Shutterstock’s library, which stands at 386 million, as reported by Interesting Engineering.
One emerging challenge in this AI-powered creative landscape is the difficulty in distinguishing AI-generated images from human-crafted ones. As more platforms incorporate generative AI into products with purchase usage rights, the potential for copyright infringement claims grows. For instance, several authors have initiated lawsuits against ChatGPT for employing their novels in its training data. Getty Images also filed a lawsuit against Stability AI for scraping millions of images from its website to train its model.
In response to this challenge, there is a growing need for an identifier that can discern AI-generated images. Google DeepMind recently unveiled SynthID, a tool capable of watermarking AI-generated images and identifying them. However, this feature is currently limited to users of its text-to-image model, Imagen.
Additionally, other tools like Maybe’s AI Art Detector have emerged, capable of determining the extent to which AI contributed to an image’s creation.
While Getty Images has not specified the exact release date for its generative AI tool on the website, they have hinted at “other service advancements” slated for later this year.
