Getty Images will outlaw the sale and uploading of AI-generated content on its platform due to uncertain copyright laws surrounding AI imagery, the company announced on Tuesday. The news follows recent decisions by websites such as Newgrounds, PurplePort, and FurAffinity, though Getty is the largest user-generated content marketplace to implement such a ban.
AI image generators such as DALL-E, Midjourney, and Stable Diffusion scrape the internet in search of publicly available pictures—often copyrighted ones on news sites, blogs, and stock photo sites like Getty Images itself—that are used to train their algorithms and which are often sampled in the images they create. Whether the purpose of using these original images qualifies as “transformative”—a major criterion of the US fair use doctrine and which most frequently entails commentary, criticism, or parody—remains unclear.
“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 told the Verge. “We are being proactive to the benefit of our customers.”
Whether Getty will be able to enforce its ban remains to be seen. To weed out AI-generated content, the company said that it would rely on users to report such images and that it would create filters with help from the Coalition for Content Provenance and Authenticity, an initiative founded last year to battle disinformation and fake news.
Daniela Braga, a member of the the White House Task Force for AI Policy, has called for a “legislative solution” for the rapidly growing and largely unregulated field of artificial intelligence.