THE MONKEY AND THE DUTCH MASTER

Who or what qualifies as an “author” under copyright law? Recent events on opposite sides of the world shed light on the question.

Last April, the Ninth Circuit Court of Appeals confronted the issue of whether animals qualify as authors. The case arose when Naruto, a 7-year old crested macaque, noticed a camera lying unattended on the jungle floor of his neighborhood on the island of Sulawesi in Indonesia. For reasons known only to himself (but which may have involved the same egocentrism harbored by humans), Naruto paused in his travels to use the camera to take several photographs of himself. These pictures, dubbed the “Monkey Selfies” by the court, went viral.

Macaquex  Next Rembrandt

While the Monkey Selfies case unfolded, a team in Holland — once the colonial master of Naruto’s homeland – were hard at work studying the facial features and proportions of the 346 known works of the Dutch Master, Rembrandt van Rijn. The team designed an Artificial Intelligence (AI) software program, incorporating a deep-learning algorithm, which analyzed Rembrandt’s use of geometry, composition, and painting materials. Utilizing a 3D printer, the   program replicated the paint and brushstrokes Rembrandt might have used, applying 13 layers of ink. It generated a new work of art, consisting of more than 148 million pixels based on 168,263 painting fragments from Rembrandt’s body of works. The portrait, known as the Next Rembrandt, was put on display in 2016, and won two Grand Prix for cyber and creative data at the Cannes Lions Festival.

The Monkey Selfies and the Next Rembrandt would seem to have little in common. But for students of copyright law, they stand together, cheek by hairy jowl, forcing reconsideration of the nature of authorship.

These issues can carry severe consequences, as David Slater, the professional photographer whose camera was temporarily expropriated by Naruto, learned to his sorrow. He became caught in a legal vice between those claiming that the Monkey Selfies are uncopyrightable and those claiming that they are copyrightable but that Naruto, not he, owns them. When he tried to stop Wikipedia from posting the Monkey Selfies on its Wikimedia Commons, Wikipedia insisted that only human beings could be copyright authors, and since the Selfies were created by an animal, they were uncopyrightable and in the public domain. When Mr. Slater included the Monkey Selfies in a self-published collection of his photographs, he was sued for copyright infringement by PETA, purporting to represent the interests of Naruto. PETA insisted that the Selfies were indeed copyrightable – just not by him. Instead, PETA argued, the copyright belonged to Naruto.  Whether Naruto desired to have his interests advanced by PETA has never been determined. In any event, PETA’s suit was rejected by the Ninth Circuit. An appeal seems likely. Meanwhile, Mr. Slater has been financially ruined and is reportedly considering a new career as a dog walker or tennis coach.

The Monkey Selfies and the Next Rembrandt raise the issue of whether copyright authorship is limited to human beings.

The Constitution empowers Congress to secure to “authors and inventors the exclusive right to their … respective writings and discoveries.” Congress has exercised that power by enacting patent and copyright statutes. Under patent law, the nexus between humans and inventorship is clear. The Patent Act defines “inventor” as “the individual … who invented or discovered the subject matter of the invention.” Only human beings qualify as inventors.

But no such clarity exists under copyright law. The Copyright Act of 1976 does not define “author.”  The Act does contain a number of provisions which suggest that only humans qualify as authors. For example, it refers to an author’s “children,” “grandchildren,” “widow,” and “widower,” all of which are generally associated with humans, not with monkeys, and certainly not with computer programs.

But the same Copyright Act also provides for two separate durations: one extends for “the life of the author” plus 70 years, while the other, for a “work made for hire,” extends to 95 years from publication or 120 years from creation. The first is for human authors. The second is for those who employ or hire humans, such as corporations and other business entities. Such entities are considered “authors” under the Act, even though they are not living, breathing human beings.

While the Copyright Act itself leaves the human element of authorship in question, those running the Copyright Office have taken a strong position. They issued the Third Edition of the Compendium of U.S. Copyright Office Practices (revised in 2017), which states: “To qualify as a work of ‘authorship’ a work must be created by a human being.” The authors of the Compendium were aware of Naruto’s Monkey Selfies and the Dutch team’s Next Rembrandt. In their examples of works the Copyright Office will not register, they included:  “A photograph taken by a monkey.” As for the computer-generated Next Rembrandt, the Compendium states: “The Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.”

The Compendium does not have the force of law. Its purpose is to guide Copyright Office employees in the performance of their duties. But its prescription is sensible and should be followed because it conforms to the fundamental purpose of copyright law.

That purpose, as the Constitution recites, is to promote the progress of “science” (which, to the framers, meant general knowledge) by granting exclusive rights for a limited time. In other words, the Constitution offers authors a deal: if they create original works, the government, in return, will reward them with a time-limited monopoly. Such incentives work only with human beings. No monkey, no matter how photogenic, and no computer program, no matter how intelligent, would be incentivized to create new works by such a deal.

The Monkey Selfies and the Next Rembrandt also lack the human “creative input or intervention” referenced in the Compendium.

Notwithstanding his visible charm, Naruto is not a human being.  David Slater, the unfortunate photographer, is a human being, but there is no evidence that he did anything more than leave his camera unattended at a propitious spot. He did not contribute the creative input required by the Compendium.

The Next Rembrandt would seem to qualify as a work of authorship on the basis of the enormous human creativity and intervention contributed by the Dutch team. But that human energy went into creating the AI program, not the painting.

Computerized Next Rembrandt

Certainly, the software should qualify as a work of authorship, and the Dutch team deserves recognition as its authors. But for the Dutch team to qualify as authors of the painting itself, they would have to show that the AI program was a mere tool, incapable of creating a work of art without human intervention.

When the only tools utilized by the creators of copyrightable works were pens or paintbrushes or chisels, the human role was clear. Such tools could not write Moby Dick, paint Whistler’s Mother, or sculpt Mount Rushmore by themselves. Human intervention was required, and humans were deemed the authors of those works.

The invention of cameras made the question of human authorship more complicated. A camera plays a larger role in the creative process than a pen, brush, or chisel does. It creates an image by mechanical and chemical functions (e.g., light passing through a lens and causing a chemical reaction on film), which operate more or less autonomously. But even with a sophisticated camera, a great deal of human creativity is still involved in the photographic process: a human being poses the subject, adjusts the shutter speed, and chooses the right film. As a result of this human intervention, no one questions the copyrightability of photographs.

The AI program created by the Dutch team is analogous to a super-camera that not only performs the mechanical and chemical functions of an ordinary camera, but goes farther. It performs all the creative functions attributed to the photographer, such as choosing the subject, setting, and film. If such a camera existed, there would be no creative function left for the human photographer to perform, and there would be no author under copyright law. Of course, there would be a patent owner for the super-camera. But his patent would not extend to the photographs taken by the device.

The Dutch team’s AI program performed all of the creative functions normally performed by the human artist. The program chose the subject, setting, colors, and brushstrokes. It left nothing for the Dutch team to contribute to the actual painting. Creating such a program is a spellbinding achievement. But that is all the Dutch team created. They are the authors of the AI computer software program. But they are not the authors of the Next Rembrandt. That work has no author under copyright law.

Annemarie Bridy, writing in the Stanford Technology Law Review, suggests that the owners of an AI software program should be deemed authors of any resultant products under the work for hire doctrine.  That doctrine separates the author-in-fact (e.g., a newspaper reporter) from the author-in-law (e.g., a newspaper company that hires and pays the reporter to write stories), and vests authorship in the latter. Under her logic, the Dutch Team could be considered an employer, and the AI program, its employee.

It’s an admirable effort to fit traditional copyright doctrines onto innovative technologies. But the fit is imperfect. In the case of real works made for hire, there is a human being — such as a reporter — doing the actual creative work. The law may not deem her the author, but she is still incentivized to create writings by virtue of the monopoly granted to her employer. That monopoly allows her employer to pay her to do her job. If the employer does not pay her enough, she may go to work for a competitor newspaper, or she may choose to freelance and thus qualify as an author herself.  Regardless of her choice, the creative process involves human creativity generated in response to the incentive system established by copyright law.

Such human creativity and responsiveness are missing from the creation of the Next Rembrandt. The AI program operated autonomously, creating a painting without human intervention, and oblivious to the incentive offered by the Constitution.

In short, the Next Rembrandt, no matter how impressive, lacks the human input or intervention required of authorship.  It is no more entitled to copyright protection as a work of authorship than a selfie taken by a curious macaque in an Indonesian jungle. We may salute both Naruto and the AI program for what they have contributed to our general happiness. But neither is human, neither did its work in response to the Constitution’s incentive — and neither is an author.

3 Comments

Filed under Law

3 responses to “THE MONKEY AND THE DUTCH MASTER

  1. Andy Strojny

    What if a corporation hires the creators of the AI program to make a Rembrandt. Are you saying if I or the AI program creators make copies of that AI created Rembrandt, the corporation has no recourse under copyright law? I realize this might be mitigated by a well drafted contract, but what about copyright law.

  2. What if a corporation hires the AI program creators to create, author, make a Rembrandt, and the AI program creators make duplicates to sell or give away. Are you saying the corporation has no remedy under copyright law. I understand a well written contract might mitigate this problem, but that’s another question. How is this different than EXXON doing much research finding a name that had absolutely no relevance to anything and than copyrighting the name or registering it as a trademark.

  3. I should add I think EXXON used a computer program to come up with the name.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s