Who Owns The Copyright Of AI-generated Content?

The time has come for lawmakers to settle this debate before it’s too late.

Some days ago, a court in Shenzhen, China ruled that a creative work produced by artificial intelligence (AI) enjoyed copyright protection. With an increase in use case of AI in content generation, be it an article or a painting, whether such work attracts the provisions of a country’s copyright act needs to be addressed. The sooner the better.

In the China case, an online platform accused of reproducing an informative article created by an AI tool without permission was asked to pay a fine for financial losses and for infringing copyright.

The question left on the table though, was — what about the copyright of those content owners whose works were used by the AI tool to generate that info article in the first place?

Before the advent of AI in content matters, copyright laws were fairly easy to understand. The thumb rule is never to reproduce works of any writer or artist without permission and/or payment of royalty/fees.

In the case of IT services, any works, including software, etc developed using licensed tools does not attract the provisions of copyright because the tools merely “support” the creative/development process. Of course, the output of this process is “copyrighted”.

Here’s an example: developers use Y platform to make a phone app, and then sell it in the market for a fee. The copyright/patent/trademark of the app rests with the developer. The fact that he used Y platform to make the app though, does not mean he has “violated” any copyright of that platform/tool. It’s akin to using a pen, a palette or a brush.

In case of creative works, so far, two factors were necessary for them to qualify for copyright protection: originality, and a human creator. With AI coming into play, a red flag has been raised on the copyright of the works it creates.

The debate so far: AI is being used to write articles, news reports; generate works of art, music even. So, there are those who argue that these should be deemed free of copyright because they have not been created by a human, nor are they “original”.

The opposition counterargues that till such time that a machine can think “creatively” on its own, without human intervention, a human being IS responsible for creating the data base. It is Man who “inputs” content for the Machine to learn. Much of the matter used for such ingestion is protected by copyright.

So…..(a) Man needs to take prior permission of the authors/creators of such inputs (b) since copyrighted material is used as input, the output, too, even though created “artificially”, should enjoy protection against unbridled shares or free reproductions.

Then, there’s one more question facing lawmakers: who should be the “owner” of the copyright? The machine, i.e. the software/hardware? Or the company behind it? Or the process? Or the Man who helped the Machine ingest the inputs?

Let me illustrate this with an example: NeoCreatives is a (hypothetical) AI company in the UK that is into creating laminated designs for table-tops, carpets and floors. For this, its chief designer Larry Southallwestminister (fictitious) uses a mix of various reproductions of some famous paintings by Leonardo Da Vinci, Pablo Picasso, and “written equations” by Albert Einstein for effect. He keeps adding more and more such works into his machine and eventually, along with the chief coder, “trains” it to turn out multiple table-top designs and carpets which are a combination of bits of the works of arts, theorems, etc. So, for every individual design, who owns the copyright then? (a) No one (b) Larry Southallwestminister © His coder (d) NeoCreatives (e) (And/ also) All those whose content/works were originally used to create the AI-based designs?

This argument around AI developed content can even be extended to patents, but let’s leave that aside for now.

Since about 2017, law makers of various countries have been looking at the issue of copyright related to AI-generated content. To my knowledge the word is still out.

Let’s look at what the United States has done so far. According to reports, recently, the US office responsible for patents and trademarks has sought public opinion on the topic. The United States Patent and Trademark Office (USPTO) has even published a notice in the Federal Register on this matter.

The USPTO has laid out 13 specific questions around the impact of AI on copyright, trademark, and other intellectual property rights. Here are some of them:

a) Should output made by AI without any creative involvement from a human qualify as a work of authorship that’s protectable by US copyright law?

b) For a company that trains an AI algo using copyrighted material for the training in the first place, should those copyrights matter?

c) Assuming involvement by a natural person is or should be required, what kind of involvement would or should be sufficient so that the work qualifies for copyright protection? For example, should it be sufficient if a person (i) designed the AI algorithm or process that created the work; (ii) contributed to the design of the algorithm or process; (iii) chose data used by the algorithm for training or otherwise; (iv) caused the AI algorithm or process to be used to yield the work….

d) How, if at all, does AI impact the need to protect databases and data sets? Are existing laws adequate to protect such data?

Some proponents of “no copyright for AI content” are trying to make a finer distinction. This group believes that if AI is used to create a “push button” output, then the outcome should not attract any copyright. Which means this — if a web developer uses a template-based tool to design/develop a website, and if he can flow AI-developed content into the AI-developed template at the push of a button, then the final website (output) should attract no copyright violation. The logic being applied here is that an AI tool eventually “figures out” for itself how to proceed and thus, creates its own design/content, etc without any human intervention at all. Push a button and the “output” comes out.

Suffice to say that like data protection laws that have finally come into play around the world (too late, if you were to ask me), new laws need to be framed around copyright issues on AI-generated content. Quick and clear decisions are required or, like in the case of online data privacy, it may become a case of too little too late.

One of the questions asked by the US patent office under its public opinion program — Are current laws for assigning liability for copyright infringement adequate to address a situation in which an AI process creates a work that infringes a copyrighted work — clearly deserves a resounding, “No” in reply.

Just think of the millions of artistes, authors and gamers whose labor of love and hard work are being used as “inputs” for AI programs to churn out machine-based content. So far, they enjoy no protection under the law, nor are they paid a fee for their content. But their contribution must be legally recognized. After all, there’s no such thing as a free lunch, right?

Image by Gerd Altmann from Pixabay