Fearing copyright issues, Getty Images bans artworks created by artificial intelligence


A selection of Stable Diffusion images crossed out.
Zoom / A selection of Stable Diffusion images with crosses through.

Ars Technica

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

To clarify the new policy, The Verge spoke with Getty Images CEO Craig Peters. “There are real copyright concerns about the output of these models and unaddressed rights issues regarding the images, the image metadata, and the individuals in the images,” Peters told Publishing.

Getty Images is a large repository of stored and archival images and illustrations, often used by publications (such as Ars Technica) to illustrate articles after paying a license fee.

Getty move Follow Image synthesis was blocked by smaller art community sites earlier this month, which found their sites inundated with AI-generated work that threatened to flood artwork created without these tools. Getty Images competitor Shutterstock Allow AI-generated artwork on his site (and though Vice I mentioned recently The site was removing the AI ​​artwork, and we’re still seeing the same amount as before – the terms for submitting Shutterstock content haven’t changed.)

Notice from Getty Images and iStock about file bans
Zoom / Notice from Getty Images and iStock about a ban on “artificial intelligence-generated content.”

Getty Images

The ability to AI-generated artwork has not been tested in court, and the ethics of using artists’ works without consent (including artworks) have found on Getty Images to train neural networks that can create artwork on a nearly human level open question It is being discussed online. To protect the company’s brand and its customers, Getty decided to avoid the issue entirely with its ban. However, Ars Technica searched the Getty Images library and found AI-generated artwork.

Can AI artwork be copyrighted?

While the creators of popular AI photomontage models insist that their products create copyrighted work, the issue of copyright over AI-generated images is not yet fully resolved. It is worth noting that the file Often cited In the Smithsonian’s “U.S. Copyright Office Rules, Artificial Intelligence Art Cannot Be Copyrighted,” it is mistitled and often misunderstood. In this case, the researcher attempted to register the AI ​​algorithm as a non-human copyright owner, which the Copyright Office denied. The copyright owner must be a human (or a group of humans, in the case of a company).

Currently, AI image synthesis companies operate under the assumption that the copyright to AI artwork can be registered to a person or company, just as it would output to any other art tool. There are some strong precedents for this, and in the copyright office 2022 decision He refused to register Amnesty International’s copyright (as noted above), and referred to a landmark 1884 legal case that confirmed the copyright status of photographs.

Early in the camera’s history, the defendant in the case (Borough Giles Lithography Co. against Saroni) He claimed that the images could not be copyrighted because the image was “a copy on paper of the exact features of some natural thing or person”. In fact, they argued, the image is the work of a machine and not a creative expression. Instead, the court ruled that the images could be copyrighted because they “represent original intellectual concepts of [an] author.”

People familiar with the AI ​​generative art process as it stands now, at least with regard to text-to-image generators, will recognize that their image compilation outputs are “representatives of the original intellectual concepts of [an] The author” too. Despite misconceptions to the contrary, creative input and human guidance are still necessary to create a photomontage work, no matter how small the contribution. Even the choice of tool and the decision to implement it is a creative act.

Under US copyright law, pressing the shutter button of the camera is randomly directed at the wall Still sets copyright For the person who took the photo, however the human creative input into a photomontage artwork can be much more comprehensive. So it would make sense for the person who started the AI-generated work to own the copyright to the image unless otherwise restricted by the license or terms of use.

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



Source link

Related Posts

Precaliga