Fearing copyright issues, Getty Images bans AI generated graphics

Increase / Selection of stable diffusion images with a cross.

Ars Technica

Getty Images has banned the sale of AI generative graphics created with image fusion models such as Stable Diffusion, DALL-E 2, and Midjourney, reports The Verge.

To explain the new policy, The Verge spoke to Craig Peters, CEO of Getty Images. “There are serious concerns about the copyright on the results of these models and unresolved rights issues with regard to the images, image metadata and persons contained in the images,” Peters said in the publication.

Getty Images is a large repository of stock and archive photos and illustrations, often used by publications (such as Ars Technica) to illustrate articles for a license fee.

Getty’s move follows the image synthesis bans introduced by smaller art community sites earlier this month, which found their sites flooded with AI-generated artwork that threatened to overwhelm artworks created without the use of these tools. Getty Images competitor Shutterstock allows AI-generated graphics on their site (and while Vice recently announced that the site was removing the artificial graphics, we’re still seeing the same amount as before – and Shutterstock’s terms for uploading content haven’t changed).

Notice from Getty Images and iStock about the ban
Increase / Getty Images and iStock Notification Banning “Artificial Intelligence Generated Content”.

Getty Images

The possibility of copyright protection for works created by artificial intelligence has not been tested in court, and the ethics of using works by artists without permission (including works found in Getty Images) to train neural networks that can create almost human works is still an open question. discussed online. To protect the company’s brand and its customers, Getty chose to avoid the ban problem entirely. That said, Ars Technica searched the Getty Images library and found art generated by artificial intelligence.

Can AI graphics be copyrighted?

While the makers of popular AI image synthesis models insist that their products create copyrighted works, the copyright issue for AI-generated images has yet to be fully resolved. It is worth noting that the frequently cited Smithsonian article titled “US Copyright Office Rules AI Art Can’t Be Copyrighted” is misunderstood and often misunderstood. In this case, the researcher tried to register the AI ​​algorithm as a non-human copyright owner, which was denied by the Copyright Office. The copyright owner must be a person (or a group of people in the case of a corporation).

Currently, AI image fusion companies operate under the assumption that the copyright of AI graphics can be registered with a human or a corporation, just like any other art tool. There is a strong precedent, and in a 2022 Copyright Office decision rejecting AI copyright registration (as mentioned above), it refers to a landmark 1884 court case that confirmed the copyright status of the photos.

At the beginning of the camera’s story, he was accused in the case of (Burrow-Giles Litographic Co. against Sarony) he argued that photographs could not be copyrighted because a photograph was “a reproduction on paper of the exact characteristics of some natural object or person”. As a result, they argued that the photo is the work of a machine and not a creative expression. Instead, the court ruled that the photos could be copyrighted because they were “representatives of original intellectual concepts. [an] author.”

Those familiar with the current AI generative art process, at least with regard to text-to-image generators, will recognize that their image synthesis results are “representatives of original intellectual concepts [an] author “. Contrary to misconceptions, human creative input and direction are still necessary to create a work of fusion of an image, no matter how small the input. Even choosing a tool and deciding to make it is a creative act.

Under US copyright law, pressing the shutter button of a camera accidentally aimed at a wall still grants copyright to the man who took the photo, yet the creative contribution of a human to a work that represents the synthesis of an image can be much broader. Thus, it would make sense for the person who initiated the AI-generated work to own the copyright of the image, unless otherwise restricted by a license or terms of use.

All that said, the copyright issue for AI works has yet to be legally resolved in one way or another in the United States. Stay tuned for further developments.

Leave a Reply