A question I have been receiving recently from readers of Educators Technology is whether AI images can be copyrighted or not. While I am not a lawyer or expert in these matters, I did some research and hopefully can help clarify some of these questions. For practical reasons, I only talk about copyright in the context of the U.S.

Is AI generated art copyrighted?

The short answer is no. AI generated artwork, according to a recent ruling by the U.S.. Copyright Office , is not protectable by copyright because it lacks ‘human authorship’.

Long answer, read below.

AI is taking the web by storm and soon AI-generated content will be everywhere online. Throughout human history, new technologies, at least in their initial phase of adoption, have always been met with resistance, skepticism, and sometimes downright refusal. This was the case with the introduction of the technology of the book, the printing press, the TV, the telephone, mobile phones, the Internet, and now generative artificial technologies.


The reasons for doubt and resistance differ from one technology to the other but at the core of this resistance, as Calestous Juma (author of Innovation and Its Enemies: Why People Resist New Technologies) argues, is that people’s “sense of what it means to be human lies at the root of some of the skepticism about technological innovation.”

In our present day, generative AI technologies and more specifically AI art generators are at the center of the storm. Reasons? Plagiarism, lack of authenticity, lack of originality, among several other reasons.

AI art generators are programs endowed with immense computing capabilities that render them capable of generating photorealistic images from text prompts.

Anyone can, in a matter of seconds, generate a ‘unique’ piece of art based on a few textual phrases or visual cues. This is all good till we take into consideration the way AI art generators work.

But in order to understand the problem of AI copyright we need to understand how these AI generators work.


AI art generators or bots have no thinking abilities and therefore can not produce original or creative work. This is at least the main argument of their detractors.

These programs are trained on a huge corpus of data from the Internet and are taught how to use natural language processing (NLP) models to learn and execute commands. For instance, the popular AI art generators Midjourney and Stability Diffisuion are trained on the open-source LAION-5B dataset which contains tons of images from all over the Internet.

So?

When you ask an AI bot a textual prompt, it simply crawls the web, synthesizes available information, and generates an answer for you. This seems acceptable at first knowing that as researchers that is exactly what we do. 


We identify our research problem, start reading into the literature by conducting various types of searches, then synthesize what we learned into a coherent body worthy of producing novel and valued contribution.

We would, however,  do all of that while referencing and recognizing those on whose shoulders we stand which is not the case with AI art generators. According to their critics, AI generators use others’ shoulders without recognition or permission.

Indeed, several established artists and cartoonists are already suing AI companies over allegations of using their artwork without permission. The cartoonist Sarah Andersen described the work of these AI art generators as “a violation of the soul”. 


Other artists view the widespread of AI generated images as “a creative invasion“. The popular image-hosting platform Getty Images has also filed a lawsuit against an AI generator accusing it of using its photos to create AI images.

Besides problems of authenticity and originality, AI art generators pose a serious threat to the livelihood of several ‘creative people’ especially artists and musicians but regardless of the motivations driving criticism and legal battles against AI generators, the question that remains to be answered is whether AI generated AI artwork is copyright free or not and if not who holds its copyright.

What does the US copyright office say about AI?

The U.S. Copyright Office has recently ruled that images generated through artificial intelligence can not be copyrighted because they lack ‘human authorship”. This verdict comes in the context of a claim made by computer scientist Stephen Thaler who applied to register a two-dimensional artwork created by his own AI program called Creativity Machine. 


Thaler’s artwork  is titled “A Recent Entrance to Paradise”. The U.S. Copyright Office refused the application to register his artwork because it lacks ‘human authorship’.

A Recent Entrance Into Paradise, by Stephen Thaler/Creativity Machine
The Office further explained that the current copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind”. Therefore, for a creative work to be copyrighted it “must be created by a human being”.

The Office added that it will not register works “produced by a machine or mere mechanical process” that operates “without any creative input or intervention from a human author”.

Bottomline, as of right now, at least in the United States, AI artwork is not protectable by copyright. However, it is important to stay up-to-date on the latest developments in this topic because things are chaning fast and new developments are taking place every single day.

Disclaimer: I am not a lawyer and I do not offer legal advice. This is a simple editorial piece that is meant to be informative. For any legal guidance on the copyright or any other legal issue, you need to seek the services of a licensed lawyer.