Ep. 19 – Who owns the Intellectual Property of an AI-based system?

In today’s Episode, we will discuss the implications of an Artificial Intelligence System building new solutions and the legal consequences on copyrights and patents. Subscribe if you want to be informed of new episodes.

As technology progresses and AI systems change roles, from doing repetitive tasks and assisting human decision-making to work with autonomy and leading an extended team comprised of Humans and other AI systems, an important question arises:

Who owns the copyrights to achievements created by AI-based entities?

Think of this: The purpose of an AI system is to make decisions and find solutions to problems via a learning process. This learning process may/will produce scenarios and solutions not envisioned by the developers of the AI system.

This week, Bloomberg Law published a piece of news where Stephen Thaler, CEO of Imagination Engines, is suing The Copyright Office, asking for granting Copyright to AI-Machines. Previously, The Copyright Office denied the claim twice because an AI machine is not a human entity.

The main dispute here is defining what a person or legal entity is – as we have natural entities (individuals) and legal entities (organizations), there is an increased interest in defining other forms of legal entities made of Intelligent systems.  

Does it make sense for an AI System to hold a patent? (or a Copyright)

I think we will not see a significant change in legislation in a short time. Still, there is a sign of changes in other geographies: Silicon Republic reported that an Australian court ruled last year that AI systems can be an inventor on patent filings.

Under the current legislation, what is subject to a patent?

In the US, the United States Patent and Trademark Office (USPTO) provides clear guidelines for this in Title 35/101. Four requirements:

1. “A” patent- means only one patent granted for each invention.

2. “Useful” – The invention must have a specific, substantial, and credible utility.

3. “Process, Machine, Manufacture, Composition of Matter” – These are the Subject matter eligibility. These categories, as interpreted by the courts, limit the subject matter that is eligible for patenting.

4. “Whoever invents or discovers” A patent may only be obtained by the person who engages in the act of inventing. This is the crucial criteria in dispute; what defines a person?

Under the current legislation, what is subject to a Copyright?

Copyright is a type of intellectual property that protects original works of authorship as soon as an author fixes the work in a tangible form of expression. In copyright law, there are a lot of different types of works, including paintings, photographs, illustrations, musical compositions, sound recordings, computer programs, books, poems, blog posts, movies, architectural works, and plays, among others. (source: U.S Copyright Office)

Same way, Copyright is granted to the owner of any original work, like taking a photograph, writing a poem or blog, or recording a new song. Everybody is a copyright owner: Companies, organizations, and other people besides the work’s creator can also be copyright owners.

I got lost; may you please provide an example?

Imagine an artist made a masterpiece: a painting, statue, or even a photograph. Does the artist deserve compensation if somebody reproduces the piece of art?

Indeed, for example, all images published here were authorized by the owner to reproduce by paying a fee.

What happens if an AI-System makes the piece of art?

Who deserves compensation for the reproduction of this piece of art?

Did you realize that an AI system made the image I published at the top of this Episode? (Dall-E2)

It was created via inputting the following command in the Dall-E2 Console:

“A palm with a tree growing on top of it”

In this specific case, who deserves compensation for reproducing the image in this Newsletter? A couple of choices to pick from:

  1.  The AI System: (Sure) This image was not previously defined or conceived by the Dall-E2 development team.
  2. Humans behind the Dall-E2 project: (Exactly) As Dall-E2 is not a legal entity, all rights are owned by the humans who lead the Dall-E2 project.
  3. The team responsible for training Dall-E2: (Hmm, that’s a good choice) Some people may argue that if the Dall-E2 masterpiece is not a consequence of a specific set of coding instructions, probably is a consequence of the training data (Machine Learning) utilized.  

Any guesses?

I don’t have a clear winner, and probably you don’t either, but I think it may be simple to recognize that the image has some value, and somebody should be compensated for reproducing the picture.

And this is a simple, trouble-free example.

Imagine if the AI-System develops a new medication, a new vaccine, or a new chemical compound to improve the efficiency of a solar cell, along with other critical use cases.

What needs to happen to resolve this singularity?

This is a space for personal opinions. My interpretation is that before we reach a new consensus on these essential topics, other aspects such as the Three Laws of Robotics and Ethical AI will be reviewed, updated, and enforced.

Section Bar

Good enough?

What are your thoughts on the subjects raised in this edition of the Digital Acceleration Newsletter?

Share them in the comments below, and if you have ideas about other topics you’d like to see covered in this Newsletter, feel free to add those suggestions as well.