Works that have fallen into the Public Domain belong to Humanity. AIs have no right to use this content. If they want to exploit creations, they just have to generate them.
State of play
Let us recall a structural obviousness which, in these turbulent times, tends to be forgotten. The vast majority, if not almost all, of our legal arsenal of protection was designed and built when the only physical medium for the transmission of knowledge, teaching, content, and entertainment was the vector of paper. There have been some developments for sound and visual works since then, but it is only since the 1970s that laws and decrees specific to digital media have appeared. And all of these regulations were designed for use only by humans. The very notion of artificial intelligence did not exist. However, despite the fact that nothing specifies or authorizes in the texts protecting authors and contents, the use by artificial machines, we act as if these works were not protected by themselves and that we could be inspired (aspired?) freely from them, to say the least.
The notion of Public Domain
Let us take the case of what we call works in the public domain. Depending on the legislation of each country, any work falls after a certain number of years (death of its last author, or expiration of copyright), in what is called this famous “public domain”. This is how the works of Maurice Leblanc are now available to everyone and the very first Walt Disney cartoons are no longer covered by copyright. This notion of public domain is fair and relevant. It is understandable that after a certain period of time, the works carried out by one human being revert to the entire global community after this certain period of exploitation. Especially when already this time exceeds decades after the disappearance of the original creators.
But when we cite this example, we forget an essential and structuring aspect, a slope on the other side of the mountain from the protective wall. This legislation, this decision, was enacted to respond to uses by individuals, biological people. They could reproduce, by the physical or digital means of their choice, works that were in the public domain. But mentioning the name of the original author. This appropriation was also allowed to generate new creations inspired by the originals.
AIs do not have Rights as such
However, we consider by default, and this is the problem, that this protective legislation, or at least the lifting of protection, is also applicable to artificial intelligence. But this is absolutely not the case. There is no reason, and in any case it is not specified in the law, that works in the public domain can be used freely by the owners of artificial intelligence. If they want content, they just have to pay the content creators. If you want to protect your castle, the worst defense is to stay at the top of its ramparts. It is the camp that attacks us that we must neutralize. There is therefore no moral, economic, social or human reason why works in the public domain can be used as they are by the owners of artificial intelligence. If there was only one organization, an International Consortium, which was responsible for spreading and disseminating all the Knowledge from around the world to everyone, then, yes, we would be in a different approach. Here, it is about slavishly copying what has been produced by others without paying anything back to the community. This attitude is not moral, but above all, it is actually completely illegal.
AIs must create content on their own
We must therefore extend the protection of works to everything that has been created by human beings and require AI to create their own content. We would then have the creation of precisely these famous new professions (creations by technological disruption) for which AI claims so much to justify its impact. Having people specialized in dissecting knowledge and reformulating it is a truly fascinating profession. And high level. We would thus truly be in a new activity. However, for the moment, on the AI side, there are only traditional positions for developers and coders that are available. While waiting for the AI to generate its own code to run in an infinite loop.
A strategic deployment
But if there is no legal and moral reason for AIs to be able to use the public domain as is, this approach must be applied in a relevant manner. One possible path would be to protect our local and European champions, if they really exist. Why not first impose this respect and this ban on American and Chinese AI and ask our “local” companies to comply over time for what we want to protect? Because we must consider that the works of the people who preceded us were dedicated to all of Humanity, and not to a few bytes on hard drives. It is high time to put humans first and as the acronym says so well, in AI there is still more artificial than real life.
So, let’s regain our sovereignty over our cultural commons by protecting them from illegitimate appropriations.
