” I argued for the idea that we own the copyright when we generate a work using AI. I saved it as a work of visual art. My certificate arrived, and I received the number and confirmation today that it was approved. “Kris Kashtanova threw everything on his Instagram account on September 21, with delight: “ My lawyer friend gave me this idea and I decided to set a precedent. »
Dawn Zarya features a character seriously resembling actress Zendaya. Indeed, AI artists use photos of celebrities, because the profusion of these on the internet provides the material necessary for the various image synthesis software. The more the AI is fed with visuals, the more accurate the rendering will be.
Dawn Zaria. AI comics.
These programs, called Steady broadcast “, Where “ Latent diffusion models “, like DALLE, launched in January 2021 by OpenIA, Nightcafe, Imagen or Midjourney, take “noise”, initially random, which can be conditioned with a text or an image. And this, before adding more “noise”, in an iterative and calculated way, until arriving at a net and expected result. In 2015, a first breakthrough in artificial intelligence resulted in the creation of a caption for an image – the translation of an image into words. Nice shot.
READ: AI draws portraits of Lord of the Rings heroes (help!)
To hang up with US copyright, the artist and his lawyer said they had ” been assistedby the AI for this graphic novel, based on machine-generated illustrations. The artist, not the AI, would have written the story of the graphic novel, designed the layout, and made the staging choices. A presentation that was certainly decisive, since the US Copyright Office has not yet officially ruled on the subject of copyright in AI works of art. It has simply, for the time being, excluded the copyright registered on an AI as an author.
To pay Dawn Zarya copyright has been verified by United Press International (UPI). The graphic novel is available for free on the AI Comic Books site.
Already on August 18, comic book author Steve Coulson revealed, via Twitter, a link to his new comic, Summer Island. We then learned that she had been entirely through Midjourney, again.
“ 40 pages of folk horror, in the tradition of mid summer and The wicker man “, at from key phrases: A small Scottish village, pagan festival, wicker man, 1973; ink line drawing, black and white comic box, contrast“. The program then generated images from several artists’ works. Many had then denounced a counterfeit or theft, ActuaBD tells us. Coulson had not disclosed the names of the authors who fueled his project.
Still on Twitter, he explained that he was in the testing phase regarding the possibilities of AI for the world of comics, this medium embodying a perfect support for his experiments.
Summer Island. Steve Coulson. Campfire.
Questions of copyright… and ethics?
From a legal point of view, the intervention of AI in the artistic creation phase questions copyright. With the works generated by an AI, the problem of the qualification of these new productions arises.
Are they works of the mind, therefore protected by law? In this case, who is the author and the right holder: the creators of the data used by the AI or the designer of the final work? Should we stop at the development of an algorithm, the fruit of the digestion of decomposed, analyzed and assimilated works, or consider that these actions constitute acts of exploitation?
From these issues arises a necessary questioning of the specific challenges of sharing and circulating data for cultural enterprises. This is why, recently, the Getty Images image bank announced a ban on the distribution and purchase of images generated by artificial intelligence, detailed Ars-Technica . The reason: the potential and dreaded claims of copyright infringement.
“There are open questions regarding the copyright of the releases of these models. And there are unresolved rights issues regarding the images and underlying metadata used to train these models. “, had indicated in particular the press release of Getty Images.
The Stable Diffusion program, for example, was fashioned from images protected by copyright and extracted directly from the web (personal artistic blogs, information sites and databases such as Getty Images). Already, some artists and photographers are criticizing artistic creation platforms for violating their own rights.
In an effort to support the work of artists, online art platforms including PurplePort, Newgrounds and FurAffinity have also banned images from artificial intelligences.
What about France?
In France, a mission report of more than 100 pages, on the legal and economic issues of artificial intelligence in the cultural sectors, was published in 2020 by the Superior Council for Literary and Artistic Property (CSPLA). In particular, it raised the question of the use of AI in culture or the status of cultural products generated by “creative” AI, whether on the status of the works that feed the AI or the qualification of the production. algorithmic.
The report considers that positive law, in a new reading of the criteria, must receive the cultural achievements of AI. This is why, if the practice reveals needs, a special right should allow an adjusted protection.
READ: AI predicts back-to-school literary award winners
Article 4 of Directive 2019/790 of the European Parliament and of the Council of 17 April 2019, on copyright and related rights in the digital single market, set up an exception “data mining», dedicated to the uses of AI. However, the directive also provides for the possible exercise of aopt outby the rights holders, which grants a return to the reservation.
As a result, the report also affirmed the need to find other solutions for access to easier use of protected content, while ensuring the defense of rights holders. He thus proposedvoluntary general licenseswhich could favor each party.
Credits: Al Comic Books