An Artificial Brushstroke: Obscure Artists Behind the Works of AI and the Copyright Dilemma

23 March 2023

The Judicial Jenga

By the 22nd century BC, the ancient Sumerian ruler Ur-Nammu designed the first law code based on casuistry and this could be considered one of the building blocks of jurisprudence. Like one of many, it constituted the basis for many jural infrastructures, and the blocks of progress have been stacked onto it as humankind evolved, which resembles the renowned physical skill game, Jenga.

After considerable time, this superimposition formed the basis for the modern jural system, and said basis was “strong” enough to add more blocks concerning various areas to create comprehensive legislation for advancing technology.

A plurality of these additional blocks relates to various aspects of intellectual property, each to speak of, but this article focuses on the historical development and current tussle of copyright law.

Attention to this matter began with the printing press, as the reproduction of texts became readily available. Yet, it took over 200 years for the first legal statute about copyright, the Statute of Anne to come into force to protect the rights of the authors.

That reproduction did not stop there, it diversified through various technological developments such as photocopiers, radios, televisions, cameras, and the like so it became challenging to keep up with the exponentially evolving technology.

Almost concurrently with these evolvements, legislations concerning intellectual properties like the Berne convention were formed on top of that Jenga tower, but what started shaking that tower are the innovations of the last decade and this Judicial Jenga could now collapse by the move of this new player, the Artificial Intelligence.

The Multi-Talented New Kid

It can write poems, compose songs, make paintings, write blogs, write codes, diagnose diseases, and many more, and it can do all of this without a break. But to understand how it does that and to understand why it is falling on the current copyright legislation like an avalanche, we should understand how AI work.

By combining large sets of data with intelligent, iterative processing algorithms learn from patterns and features in the data that it analyzes. Each time an AI system runs a round of data processing, it tests and measures its own performance and develops additional expertise. Because AI never needs a break, it can run any number of tasks extremely quickly, learning and recording every single piece of information in a very little time, and becoming extremely capable at whatever it’s being trained to accomplish.

As this article focuses on the artist side of AI, there are two separate but dependent topics that concern AI-generated visual art.

The visuals created by the AI through an input of a human user

Currently, there are a handful of platforms used to make AI create art for you, and three of the most popular are Midjourney, Dall-E, and Stable Diffusion. Using these, you can type a scene in your mind, like “Tom and Jerry playing nine-ball on Jupiter”, and this will result in an image that looks like the thing close to what’s in your mind.

As we discuss the ownership of an AI-generated art piece, there are several ones that granted the owner not a copyright but a prize for the work, and one of the most renowned examples is Jason M. Allen’s below-shown piece, the Théâtre D’opéra Spatial.

As may it be so hard to believe, no human moved a muscle to create that image but Midjourney did, with the imagination of Allen. According to Ryan Meyer of Dorsey & Whitney LLP, “AI cannot create art, but it usually mimics or even contains reproductions of other people’s artwork to create new artwork. That new artwork could be an unauthorized derivative, which is an infringement. If the AI also stores a reproduction of that artwork, that too is an infringement. Another issue is that, while the artwork is generally in the domain of copyright law, computer technology necessary for AI-generated artwork might be protectable under patent law. We might start seeing patent law used in new ways to protect AIs and the methods they use for creating artwork”.

The US Copyright Office has already refused to grant copyright to an owner for AI-generated art as the current copyright law requires human authorship for copyright protection. This was the case of Stephen Thaler, who is known for the Dabus Case which concerned the patent application for an AI-generated invention. Said case was about recording the AI as the inventor, and this latter case was about recording the AI as the copyright owner of the below-shown piece.

That means under the current rules, AI-generated art has no owner. That creates the first burden on the current legislation, as the AI-generated art has no owner but in this case, the code has an owner so it is still relatively possible to trace the property owner of such pieces.

It gets much more complicated when the idea of the creation of general artificial intelligence, which roughly means AI that is as smart or smarter than humans. Amongst many theoreticians on this topic, Jeff Clune of the University of British Columbia suggests the idea of AI Generating Algorithms – namely AI-GA: open-ended algorithms that can learn forever. According to Clune, through an AI-GA, it is possible to create a general AI. Artificial general intelligence (AGI) is defined as the intelligence of machines that allows them to comprehend, learn, and perform intellectual tasks much like humans. This means an AI intelligent enough could both code and execute the code to do tasks like a painting, for example, as it concerns the subject matter.

This brings the second and much topsy-turvy burden on the current copyright law: an untraceable set of codes that works on its own without stopping. Imagine a code creating numerous pieces of content instantly every day and releasing it on the internet. That means thousands, or even millions of “art” wandering all around the world, infringing thousands of works and stealing artists’ ideas without a break.

In this consumption era, this is hard for both the artists and the art lovers but for now, as this kind of self-operating AI seems to be a little far ahead and as the IP regulations seem to barely catch the current evolvements, that second scenario could be procrastinated.

As stated above, US jurisdiction is still solid against this matter, as are Spain and Germany, but apparently, these won’t last long enough.

Artificial, yet intelligent to bring forward brain teasers

Up to the present, people have created things that didn’t require comprehension of the communities. Yet, they were all protectable by law. This time, the IP field deals with vagueness, and questions bring questions. For instance, even if it becomes possible to properly protect, would or should an artwork generated by AI be worthy of copyright protection?

If so, should the copyright be held by the machine itself, or the developer? Even more assertively, should it be a public right?

In this article, although the current status of the legislation and theories for the concept of entitlement to the existing Artificial Intelligence technology and the works produced by it is examined, the subject remains ambiguous, and perhaps it may be years or even centuries from now, as the Statute of Anne comes into force, for a clarification on this matter.

As Thaler is persistent in procuring acceptance for the works of AI, the AI itself shall be much more persistent in the near future and we will barely spectate.

Sources: