A work of art generated by artificial intelligence raises many questions, especially about its ownership. Who is responsible for the copyright to these works? The creator of the algorithm? The person who pressed the button to start the program? The AI itself? Can a work created by AI be considered the product of a human spirit?
Before the latest innovations in AI technologies, copyright ownership of computer-generated works was not at issue, because machines were only tools that the artist uses to create.
In most European jurisdictions, copyright protection requires a human author. Therefore, a problem arises when a machine decides and creates for humans, because these works can easily be considered copyright-free, and thus be accessible and free for all. So why would someone invest in something that they would not benefit from the end result? This property problem is commercially very bad for developers. There are a few legal options around the world, including two: either copyright protection is denied for works created by AI technologies, or it is attributed to the creator of the program. Copyright ownership to the creator of the program has never been specifically prohibited because AI is a fairly new technology and laws require some time before adapting to the reality of the world.
In the United States, the copyright office only records a work if the work was created by a human being because copyright, a branch of intellectual property law, protects only the fruits of intellectual work, which can only be carried out by the human mind. It is for exactly the same reason that J.R. Boronali, a donkey who painted the Sunset on the Adriatic with a brush attached to his tail, never had the ownership of the copyright of his painting. The Court of Justice of the European Union specifies that an original work must reflect the personality of the author, it must bear the intellectual imprint of its author.
On the other hand, in the United Kingdom and several other countries, copyright ownership is given to the programmer because it was the programmer who took the necessary initiatives to achieve the final result, the work of art. This makes sense in the case of AARON, created by Harold Cohen since he wrote the algorithm; he is also the only one using the program, but what about GAN and CAN? The Paris-based Obvious collective used GAN technology to create the famous Portrait of Edmond Belamy; does the work belong to GAN developer Ian Goodfellow, or does it belong to the person who pressed the button to launch the program? We could draw a parallel with Adobe, questioning the authorship of all the works created on Illustrator. The nuance lies in the existence of AI, because the contribution of users is so small that it could be considered non-existent and in this case, the copyright would belong to the person who created the algorithm. In reality it doesn’t really make sense…
For the time being, until legislators formulate a clear answer to this question and codify it into law, the best answer is to deal with copyright on a case-by-case basis.