Increase / A selection of images of stable diffusion with a dashed line.

Technique Ars

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. This is reported by The Verge.

To explain the new policy, The Verge spoke with Getty Images CEO Craig Peters. “There are real copyright issues with the results of these models and unresolved rights issues with the images, the image metadata, and the individuals contained in the images,” Peters told the publication.

Getty Images is a large repository of stock and archival photos and illustrations that are often used by publications (such as Ars Technica) to illustrate articles after paying a licensing fee.

Getty move follows Earlier this month, small art community sites banned image synthesis after their sites were flooded with AI-generated work that threatened to overwhelm artwork created without the use of these tools. Getty Images Competitor Shutterstock allows Illustrations created by artificial intelligence on their website (and although Vice recently reported the site removed the AI ​​artwork, we still see the same amount as before, and Shutterstock’s terms of service have not changed).

A notice from Getty Images and iStock about the ban
Increase / Notice from Getty Images and iStock banning “artificial intelligence-generated content.”

Getty Images

The possibility of retaining copyrights in works of art created by artificial intelligence has not been tested in court, and the ethicality of using works of artists without consent (including works of art found on Getty Images) for training neural networks that can produce near-human-level artwork, still open question discussed on the Internet. To protect the company’s brand and its customers, Getty chose to avoid the issue altogether by banning it. However, Ars Technica searched the Getty Images library and found AI-generated artwork.

Can a work of artificial intelligence be copyrighted?

Although the creators of popular AI image synthesis models insist that their products create copyrightable works, the issue of copyright in AI-generated images has not yet been fully resolved. It should be noted that an an oft-cited article in the Smithsonian’s section titled “US Copyright Office Rules AI Art Can’t Be Copyrighted” is mistitled and often misunderstood. In this case, the researcher tried to register the AI ​​algorithm as a non-human copyright holder, which the Copyright Office refused. The copyright owner must be a person (or a group of people, in the case of a corporation).

Currently, AI image synthesis firms operate under the assumption that AI artwork can be copyrighted to an individual or corporation, just like any other artistic tool. There is strong precedent for this, and the Copyright Office 2022 decision in rejecting an artificial intelligence copyright registry (as mentioned above), he cited a landmark 1884 court case that upheld the copyright status of photographs.

At the beginning of the camera story, the defendant in the case (Burrow-Giles Lithographic Co. against Sarona) argued that photographs cannot be copyrighted because a photograph is a “reproduction on paper of the exact features of some natural object or some person.” In essence, they argued that photography is the work of a machine, not creative expression. Instead, the court ruled that photographs could be copyrighted because they were “representative of original intellectual concepts [an] author”.

People familiar with the generative art process of artificial intelligence in its current form, at least with regard to text-to-image generators, acknowledge that their image synthesis results are “representative of the original intelligent concepts [an] “author” too. Despite misconceptions to the contrary, human creative input and guidance is still necessary to produce a work of image synthesis, no matter how small the contribution. Even choosing a tool and deciding to do it is a creative act.

Under US copyright law, pressing the shutter button on a camera randomly pointed at a wall still transfers copyright to the person who took the picture, and yet the person’s creative contribution to the work of image synthesis can be much broader. Therefore, it would make sense for the person who initiated the AI-generated work to own the copyright in the image, unless otherwise restricted by license or terms of use.

All that said, the issue of copyright in AI artwork has yet to be settled legally one way or the other in the United States. Stay tuned for further developments.

https://arstechnica.com/?p=1883513

Previous articleParis Hilton ‘desperate’ to find missing dog Diamond Baby – NBC10 Philadelphia
Next articleFox News veteran Trace Gallagher gets night anchor job | Health-med-fit