Thanks to artificial intelligence (AI), it is now possible to compose a song or paint a picture with just one click. But behind these high-tech achievements, there are complex legal issues. Who owns the rights to a work created by an AI algorithm? What does AI “learn” based on? 

In the UK, works created by AI are copyrighted, albeit for a shorter period of time than human creations (50 years from the moment the work was created instead of the author's entire life and another 70 years from the moment of his death).

The EU copyright law has established an approach in which the legal protection of a work as a result of intellectual activity is related to the presence of the result of an intellectual creation created by human being. The key principle is enshrined in the sentences of the ECJ in the cases of Infopaq, Painter and Cofemel: a work is subject to legal protection if it reflects the free and creative personal choice of the author. Thus, originality coincides with the presence of a recognizable “personal trace” of the author.

It is easy to see that this approach creates numerous questions. Should human input be structured? Is it about defining parameters, designing the creation process, or selecting and editing the results?

At the moment, law enforcement practice on these issues is in the process of formation. One of the few decisions was made by a court in Prague (Czech Republic), in this decision the court refused to legally protect an image created entirely by AI, citing the lack of independent human input.

The question of when a human contribution is sufficient for a work to be subject to legal protection remains open.

Obviously, Russian courts will have to deal with similar issues. In this regard, individuals who use AI to create works should be ready to the relevant legal risks.