Artificial intelligence is progressing every day. Today, it produces images, music or texts with a degree of autonomy beyond the stage of computer-assisted creations, which raises questions about copyright. Interview with Nathaly Vermette, lawyer in business law and intellectual property at DS Avocats.

If we turn to pure copyright, I always like to go back to the sources. The goal was to protect and encourage creativity in useful sciences by giving creators the exclusive right to share their works. The other element: there are criteria of originality, authorship and succession of the work. Copyright persists for the life of the author and 70 years after his or her death. So we can deduce that it must concern a living being.

We are left to define the quantity and quality of human intervention. There are works that are assisted in the choice of data, and there may be adjustments afterwards, for example.

In Canada, when you file a copyright, the application is simply analyzed to verify that all the elements are there. It is not examined in depth. It happens when someone sues someone else for infringement. We haven’t had a case yet, but researchers are interested in it. There are doctoral theses being written and which will be published on the issue.

In the United States, there is not yet any case law on the precise question of the amount of human input. In 2015, an American animal rights organization tried to have the copyright of a selfie of a monkey recognized by the courts. The photographer had set up his equipment and the monkey pressed the button then took the photo of himself. The Court of Appeal did not give it to him. And if it can’t be granted to a monkey, I think even less of a machine.

There are studies and groups looking into the issue. They say all AI-assisted work should qualify. Then there are those who say that all autonomously generated work should be in the public domain.

Artificial intelligence will always be able to imitate humans, but will not be human. If the legislator wants to encourage creativity, it would go against its objective of allowing artificial intelligence to hold copyright.

The works can be classified into two categories. Those which are purely human or with AI contribution, then those which are completely generated by AI. I believe that these, perhaps wrongly, will not have copyright.

There will have to be an international consensus on how to deal with this. Like a contract with member countries to achieve similar functioning. It is certain that there are several factors that we will have to determine.

If the court is unsure how to apply the existing law, the government will need to clarify it by passing amendments. This is how the law works. In Canada, as a general rule, we are always behind. We often tend to follow the United States, which is faster, and we adjust according to our laws and customs.