It was a demanding month for Midjourney. This week, the generative startup of the AI released Its new sophisticated video tool, V1, which allows users to create short clip animated by the images that generate or load. The current version of the Midjourney’s video tool requires an image as a starting point; The generation of video using text prompts is not supported.
The release of V1 arrives in the wake of a very different type of announcement in early June: Hollywood Behemoth Disney and Universal have filed a successful cause of success against Midjourney, claiming that it violates the copyright law generating images with the intellectual property of the studies.
Midjourney did not immediately respond to requests for comment. Disney and Universal reiterated the statements made by its managers on the case, including Disney’s legal head Horacio Gutierrez claiming that the production of Midjourney is equivalent to “piracy”.
It seems that Midjourney may have attempted to stage some specific guardrails for video for V1. In our tests, he blocked the animations from Elsa -based instructions, Boss Baby, Frozen and Frozen Mickey Mouse, although he would still have generated images of these characters. When Wired he asked V1 to animate Elsa’s images, a “moderator Ai” blocked the prompt from the generation of video. “Moderation is cautious with realistic videos, especially people”, read the pop-up message.
These limitations, which seem to be guardrail, are incomplete. The wired tests show that V1 will generate animated clips of a wide variety of universal and Disney characters, including Homer Simpson, Shrek, Minions, Deadpool, Star Wars‘C-3po and Darth Vader. For example, when a minion image was asked that eat a banana, Midjourney generated four outputs with recognizable versions of the nice yellow characters. So when Wired click on the “animated” button on one of the outings, Midjourney generated a follow-up video with the characters who ate a banana: beaten and everything else.
Although Midjourney seems to have blocked some Disney and Universal instructions for videos, Wired sometimes it could evade potential guardrail during tests using spelling variations or repeating the prompt. Midjourney also allows users to provide a prompt to inform animation; Using that function, Wired has been able to generate clips of copyright protected characters that behave in adult ways, such as wall-e who branded a firearm and Yoda who smoke an articulation.
The Disney and Universal cause represents a serious threat to Midjourney, which must also face further legal challenges from visual artists who also support the violation of copyright. Although it has focused largely on the supply of examples from the tools of generation of Midjourney’s images, the report claims that the video “improves only the ability of Midjourney to distribute copies, reproductions and derivatives of protected works by the sued”.
The complaint includes dozens of alleged images of Midjourney showing universal and Disney characters. The set was initially produced as part of a relationship on the so -called “visual plagiarism problem” by Midjourney from the artificial intelligence critic and cognitive scientist Gary Marcus and the visual artist Reid Southen.
“Reid and I stressed this problem 18 months ago, and there were very few progress and very few changes,” says Marcus. “We still have the same situation as non -license materials used and guardrail that work a little, but not very well. For all speeches on exponential progress in artificial intelligence, what we are getting is a better graphics, not a solution of a fundamental principle to this problem.”
Discover more from Gautam Kalal
Subscribe to get the latest posts sent to your email.
Be First to Comment