Copyright and AI in 2025: An Artist’s Perspective
As the advancements in AI and ML continue, the tech industry is trying to pick a fight they don’t deserve to win.
Defining the Issue
The best way to set the parameters for discussion is to answer this question as clearly as possible: “What is AI generated art?” Conveniently enough, AI itself has an answer this morning, Wednesday March 19, 2025, courtesy of Google:
The Origin of US Copyright and Current Protections
As many of the US Founding Fathers were prolific writers — among them Benjamin Franklin — protections for their works were documented in the Constitution:
According to Article I, Section 8, Clause 8 of the US Constitution, “the Congress shall have power . . . to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.”
For decades, the US Copyright system has been subject to legislative changes that have extended the coverage for protection. In 1790, the term was set at the life of the artist plus 14 years (with an option to renew for another 14), that period has grown to a staggering 95 years. This is frequently called “The Mickey Mouse Rule” because Disney has traditionally been the primary lobbying firm behind the changes in order to keep Mickey Mouse from entering the Public Domain.
As I understand it, the concept of Public Domain works is grounded in the original intent behind the phrasing in the Constitution: society benefits when artists are allowed to build upon prior works, as copyright enables a creator to profit during their lifetime from the protections granted, and once they are dead, their works eventually are placed in Public Domain.
The concept of a 95 year protection is a distortion of this original intent and clearly a modern movement for commercial monopolization much, much longer than the Constitution considered. Hence the easy-to-recognize motivation of protection extensions for the benefit of Disney and other rights holders (estates of dead artists, record labels, and the like).
About Fair Use
There is a specific carve-out in the US Copyright protections aimed at promoting useful progress in the arts by way of “Fair Use,” spelled out clearly for Judges to weigh during a case:
- the purpose and character of the use
- the nature of the copyrighted work
- the amount and substantiality of the portion taken, and
- the effect of the use upon the potential market.
These are very clear terms that, while legally minded, are simple enough to grasp conceptually.
From my observations, the tech industry has a very warped view of Fair Use and wave it around like a flag at a protest — when they are caught violating the copyright of artists, they whine that what they are doing constitutes Fair Use…and if it doesn’t, it should!
The tech industry is petulant in this regard. Time and time again startups have played fast and loose with copyrighted works. In several instances, piracy has been part of the early business model. Napster tried and failed. YouTube tried and succeeded. Spotify tried and succeeded. By the time rights holders felt compelled to go after YouTube and Spotify seriously, the companies were big enough to negotiate directly and both sides avoided costly litigation. The gamble paid off.
Copyright is for Humans
A recent court ruling regarding a case in 2018 has made the news and it brought joy into my heart:
Recall the AI provided definition above provided by AI:
…algorithms learn from data inputs and generate new, original art pieces…
Clearly AI knows it’s not a human, it knows it’s a bunch of algorithms. As such, when it claims to “generate new, original art pieces” that is something its designers have programmed into the algorithms. Basically, AI firms are cheating because they know full well the only way an AI system can generate content is by training it on existing content. A more truthful statement would crush their ability to defend their claims in court, so I’ll go ahead and write it out for sanity’s sake:
…algorithms learn from art created by humans and generate derivative works…
The reality in this case is that AI firms are highly secretive about what content they are using to train their systems. Going forward, this may be a significant advantage in court for protecting the concept of copyright:
Unless an AI system has been trained exclusively on works from the Public Domain, then by definition, every work generated by that system is a derivative of copyright protected material, and the resulting works can not pass the Fair Use four-factor test…
Therefore the only entities entitled to hold the rights for those AI generated works are the authors of the materials used in the training of the AI system.
This undercuts the business model of generative AI for writing creative works, for music, for films, and I’m firm in my perspective that this is the correct interpretation of copyright in the modern age. Artists are, in the US, a traditionally exploited class and economically disadvantaged, save for a select 1% who make incredible sums of money.
As an example, there’s enough Taylor Swift music to set up an AI system, train it on her material, and start cranking out songs (rather than the obscene Deep Fake content being circulated apparently). Ideally, these systems should be used to demonstrate how some of the most financially successful entertainers aren’t actually very special and their art is clearly manufactured for mass consumption. Oh, my apologies to Max Martin in advance, but then again, that ABBA-inspired hit machine is basically a walking human pop tune algorithm.
To compare, what would the AI generated music sound like if it was trained on Radiohead? Could the algorithm even begin to comprehend the difference between “OK Computer” and “Kid A” and understand this is the same band, the same musicians? This is the kind of experiment that actual human being musicians undertake frequently…
In ideal circumstances, we listen, we look, we study, we learn, and we create new art and music and writings inspired by our influences.
The recent court ruling gives me hope that major industry players — Record Labels, Film Studios — will simply see AI generated content as a financial liability. Their pursuit of revenue is not about art really, so avoiding expensive litigation is in their best interest. For the time being, no matter how much large AI firms want to assert that their systems produce original content — a dubious claim — that deserves copyright protection — clearly losing the battle in court — people in the creative industries will understand the reality…only humans create art eligible for copyright protection.
