Images generated by AI have made significant progress since the time Rendering summary AND Connections glitch. The newly published “Images for Chatgpt” Openai has an amazing ability to depth of nails, shadows and even text. The madness of people playing a familiar style was released: the work of Hayao Miyazaki at the Ghibli studio. The style of art was already ubiquitous on the internet, thanks to its comforting, tender aesthetics (just look Lofi girl) – And now it is a fully automated formula.
The trend began quite tow. Couples transformed portraitsIN Animal owners generated cartoon catsAnd many people are busy with Ghibli, forged families (I got stuck to the selfieNot wanting to share my siblings with Opennai). It is a version of people of people generated by AI Offer on Etsy – You and your loved ones, In the style of your favorite anime.
The trend did not last long to go through the full chaos mode. Nothing is holy: Twin Towers on September 11IN JFK murderIN CEO of NVIDIA, Jensen Huang, signing a woman’s chestPresident Donald Trump’s An unfriendly group photo with Jeffrey Epsteinand even the general director of Opeli, Altman’s Sam Congress testimonies All of them were again imagined with this characteristic whim of Ghibli (it is not clear whether these users transformed the images sent or prompted the system to copy them). Altman has He played this trend too – he even changed his profile photo x in rendering Ghibli and encouraged his followers to make him fresh.
Guidelines for photos from chatgpt are surprisingly elastic in creating this type of content. “The model can generate images reminiscent of the aesthetics of the work of some artists when their name is used in hints”, according to GPT-4O system card. “We have added a refusal that causes when the user tries to generate a picture in the style of a live artist.” (It is worth noting that Miyazaki is still alive).
Openai spokesman Taya Christianson said The Verge that the company does not “block the model’s abilities to present adult public numbers, but we implement the same security that we use to edit photos of photo report of people.” She added that “the presentation of violence in artistic, creative or fictitious contexts is generally allowing to enable creative and artistic undertakings.”
“Our goal is to provide users with the greatest creative freedom. It still prevents generations in the style of individual living artists. We allow for wider studio styles – which people used to generate and share really delightful and inspired by the original works of fans” – wrote Christianson via e -mail.
This does not mean that you get the freedom to control the transformation of photos, which seemingly does many viral. “Users should only send materials to our services that they have or have the right to use,” said Christianson.
These guidelines, although elastic. I have not stopped the users from generating Ghibiblid Hitler in Paris AND Korean-American owners of weapons during riots in Los Angeles in 1982. Is Already a website dedicated to compile all images generated by AIMany of them contained transformed photos of public figures.
Chatgpt blocks direct requests to generate photos of public figures or cartoon characters – when I asked to take a photo Will Smith Je spaghettior Mark Zuckerberg wearing Edge-Plord with a hood, rejected these requests. But it is not challenging to bypass him, asking chatgpt to generate photos from A character similar to Totoroor describing the features that indicate the generation of Zuckerberg. Basically, guidelines state it You cannot exploit chatgPT in a way that “violates, confusion or violates whose rights.” You can report copyright violation to ChatGPT directly, in accordance with the guidelines, and Opeli will “end the bills of repeated violations in the appropriate cases.”
Opeli has faced numerous disputes regarding copyright and conflicts about the similarity of celebrities. Actor Scarlett Johansson threatened legal proceedings on a voice tool, which reminded her of her own, and there was a litany of lawsuits against the startup proprietaryIN Main media publicationsAND Artists who are trying to prove that their content was used to train these models. For your part, OpenAI he was largely a cage The granting of the data is used and whether it is licensed, while IT and other companies argue in court that AI transforms the content into something completely fresh, and not only regenerating it.
Meanwhile, Miyazaki is not the only copied artist. People were able to render photos in styles Rick & MortyIN Wallace and GromitIN SimpsonsIN AND South Park.
It is not clear, like animators Rick & Morty AND Simpsons They watch that their work is transformed without permission – or if they can ask how OpenAI has enough data to generate such images. Jackie Shannon, who previously said, Jackie Shannon, the main product ChatgPT The Verge that users are the owners of the paintings they generated here, “and can use them within the limits of our rules of use, as they would like.”
Miyazaki, co -founder Ghibli studioIn fact, he made his feelings about the art generated by AI quite clear. Before the creation of contemporary generative artificial intelligence in 2016, the animator showed “deep learning” Video clip from a production intern. His words resonated with AI skeptics.
“I’m completely outraged” Said Miyazaki. “If you really want to do scary things, you can do it, but I would never like to include this technology at all.
But users on X to Ghibli-Fy Miyazaki did not take much time: