“The criterion is not whether Kashtanova has ‘a touch of creativity’ rather than ‘what can you anticipate along the way? [would] Please create it in advance.In other words, the office is wrongly focused Tool output not human input“
The authors, along with Max Sills of Open Advisory Services, represent Kashtanova in the cases discussed below.
The United States Copyright Office (USCO) announced its decision last week in Kristina Kashtanova’s comic book lawsuit. Dawn of ZaryaKashtanova retains copyright registration, but it is limited to the entire work as text and compilation.
In some ways this is a success. Because the office had previously threatened to revoke the copyright entirely. However, the Secretariat has restricted registration and specifically excluded individual images created by Kashtanova from copyrighted material. This uses artificial intelligence (AI) tools as part of the creative process. It’s a step back for any artist who wants to.
Given the context of other AI-driven cases the office is dealing with, this is a surprising result. In his recent USCO filing on the Stephen Thaler case, the agency said it was “preparing guidance for registration of works generated using AI.” If the Secretariat is preparing guidance permitting the registration of AI-assisted works, it strongly suggests that the USCO believes it exists. Several Threshold of human involvement sufficient to allow enrollment.
The Bureau is aware of Kashtanova’s input into the images created for Zarya of the Dawn, but claims that human input has had no material effect on the resulting images. It is possible to predict in advance what Midjourney will create” (p. 8) or “Ms. Generate an image.” (p.9) [Editor’s note: The USCO’s letter incorrectly refers to Kashtanova using she/her pronouns; Kashtanova’s pronouns are they/them].
A little input is enough
This is not the correct legal standard.It’s not about what Kashtanova can anticipate on her journey, but about whether she has “a little creativity” [would] Please create it in advance. ” In other words, the office is wrongly focused Tool output not human input.
Jackson Pollack famously failed to predict how the paint he used dripped onto his canvas. Pollack designed his paintings, but he used a process of randomly dripping and plucking paint to create his art.In the music, each performance of John Cage’s 4 minutes 33 seconds fully defined By random sounds made by the audience and the world around the stage.
Photography, the closest analogical art, has famous photographs that completely misrepresent animals, people, or even humorous situations. Nonetheless, the output is still copyrighted because humans have at least “minor” input to the shot.
Of even greater concern to AI artists is the misunderstanding of the facts that drive this decision. The Secretariat seems to believe that a defining characteristic of AI tools is that the output is ‘random’ or ‘unpredictable’. As the Secretariat states, “Midjourney differs for copyright purposes from other tools used by artists due to the fact that users cannot predict the specific output of Midjourney.” (p.9) )
of just The output may be unpredictable, but experiments have shown that typing “cute purple dinosaur” into Midjourney returns images of cute purple dinosaurs, not motorcycles or clouds. Moreover, the more input the artist gives, the more control over the output. Artists can design their output to have a specific subject matter, lighting, content, layout, and atmosphere. It doesn’t matter that subject matter, lighting, content, and layout are generated rather than captured.
AI-assisted art is still in its early stages, and USCO participates in many outreach events to get feedback. Ultimately, however, AI-powered art must hold to the same standards as photography, with enough human input in it to warrant copyright protection, even for rude photographs.
Image Source: Deposit Photo
Image ID: 31248541
Author: Tashatuvango
