"Early generative AI technologies have raised questions about how it should be properly used," Sarah Casillas, senior director of content for Adobe Stock," wrote in a blog post. "Adobe has deeply considered these questions and implemented a new submission policy that we believe will ensure our content uses AI technology responsibly by creators and customers alike."
AI-generated art has quickly captured the attention of the public, with fun and buzzy apps like DreamUp and AI Time Machine giving users original and compelling illustrations based on the users' own pictures or text prompts. However, the technology -- based on generative AI models like Dall-E and Stable Diffusion, which have been trained on content scraped from the internet -- has raised several questions about the nature of art, as well as legal questions such as how artists should be compensated.
Casillas called generative AI "a major leap forward for creators", giving them a way to expand their portfolios and their earning potential. She added that Adobe intends to help turn generative AI technology into "tools that empower artists, while never seeking to replace human imagination."
To ensure that customers on the Adobe Stock platform know what they're paying for, Adobe's new generative AI policy requires contributors to label their AI-generated content as such. It also requires contributors to get the proper authorization before submitting art based on third-party content -- such as text prompts referring to people, places, property, or an actual artist's style.
While Adobe is trying to embrace generative AI, some other content platforms have moved in other directions, restricting or banning AI-generated content.
"There are real concerns with respect to the copyright of outputs from these models and unaddressed rights issues with respect to the imagery, the image metadata and those individuals contained within the imagery," Getty Images CEO Craig Peters said to The Verge.
Meanwhile, some online art communities have banned AI-generated images, concerned their platforms could be overwhelmed with the easily produced content.