Afraid of copyright issues, Getty Images bans AI-generated artwork

Enlarge / A selection of crossed out stable broadcast images.

Ars-Technica

Getty Images has banned the sale of generative AI artwork created using image synthesis models such as Stable Diffusion, DALL-E 2 and Midjourney through its service, reports The Verge.

To clarify the new policy, The Verge spoke with Getty Images CEO Craig Peters. “There are genuine copyright concerns about the outputs of these models and unresolved rights issues regarding the imagery, image metadata and the people contained in the imagery,” Peters told the publication.

Getty Images is an extensive repository of stock and archival photography and illustrations, often used by publications (such as Ars Technica) to illustrate articles after paying licensing fees.

Getty’s move follows synthetic imagery bans imposed by smaller art community sites earlier this month, which found their sites inundated with AI-generated work that threatened to overwhelm created artwork. without the use of these tools. Getty Images competitor Shutterstock allows AI-generated artwork on its site (and although Vice recently reported that the site was removing AI artwork, we’re still seeing the same amount as before – and the Shutterstock’s content submission requirements have not changed).

A notice from Getty Images and iStock on the prohibition of
Enlarge / A notice from Getty Images and iStock on banning “AI-generated content”.

Getty Images

The ability to protect AI-generated artwork has not been tested in court, and the ethics of using artists’ work without consent (including artwork found on Getty Images ) to train neural networks capable of creating near-human works of art is still an open question. debated online. To protect the company’s brand and its customers, Getty decided to sidestep the issue entirely with its ban. That said, Ars Technica dug into the Getty Images library and found some AI-generated artwork.

Can AI artwork be copyrighted?

While the creators of popular AI synthesis models insist that their products create copyrighted works, the issue of copyright on AI-generated images has not yet been fully resolved. It should be pointed out that an oft-cited article in the Smithsonian titled “US Copyright Office Rules AI Art Can’t Be Copyrighted” has the wrong title and is often misunderstood. In this case, a researcher attempted to register an AI algorithm as a non-human copyright owner, which the Copyright Office denied. The copyright holder must be human (or a group of humans, in the case of a corporation).

Currently, AI synthesis companies operate under the assumption that copyright for AI artwork can be registered in the name of a human or a company, as is the case for the output of any other artistic tool. There is strong precedent for this, and in the Copyright Office’s 2022 decision denying the copyright register to an AI (as mentioned above), it was referring to a landmark legal case from 1884 which affirmed the copyright status of the photographs.

Early in the camera story, the defendant in the case (Burrow-Giles Lithographic Co. c. Sarony) asserted that photographs could not be copyrighted because a photograph is “a reproduction on paper of the exact characteristics of a natural object or person.” Indeed, they argued that a photo is the work of a machine and not a creative expression. Instead, the court ruled that the photos may be copyrightable because they are “representative of original intellectual designs by [an] author.”

People familiar with the AI ​​generative art process as it currently stands, at least as far as text-to-image generators are concerned, will recognize that their computer-generated image outputs are “representative of intellectual conceptions original by [an] author” as well. Despite misconceptions to the contrary, a human’s creative input and guidance is always needed to create computer-generated imagery work, regardless of the input. Even the tool selection and decision to performing it is a creative act.

Under US copyright law, pressing the shutter button on a camera randomly pointed at a wall still assigns copyright to the human who took the photo, and however, the human creative contribution in an image synthesis work can be much more extensive. It would therefore make sense for the person who initiated the AI-generated work to own the copyright in the image, unless otherwise restricted by a license or terms of use.

That said, the issue of copyright in AI artworks has yet to be legally resolved one way or another in the United States. Stay tuned for future developments.

Comments are closed.