.png)
It was January 2016 when in Seoul the world champion of the board game Go, Lee Sedol, was defeated live by the artificial intelligence system developed by Google's Deepmind team, AlphaGo. Many recognize that event as the revelation to the world of the immense potential developed by Deep Learning systems, the technology that applies to artificial intelligence.
Less than 10 years since that day, giant strides have been made: many companies have begun to engage AI models to carry out their activities, and this technology has pervaded so many industries, becoming the most effective system for improving and making our world more efficient. There is more and more talk about this technological development, its advances and risks, and today there are even those who express great concern about the little regulation and unprecedented speed of this development.
But what is meant by Deep Learning? It involves providing the computer with the rules that govern a certain area, for example, the rules of the road, the rules that govern the evolution of a disease or the rules of a game, as in the case of Google to which we referred in the opening, and at the same time a giant number of "cases" that have occurred in the past in that area. Using an algorithm, the computer begins to participate in its activity, initially with poor or even disastrous results, and then, slowly, it begins to understand the rules of the game and participate successfully, even developing new and innovative strategies that had never been pursued before.
The revolution brought by AI systems has not left indifferent even the artistic or humanistic sectors - let's note how the word acquires even more meaning in this context - which have recently brought texts, images, photographs and other material generated by software onto the public scene of AI. In the same year in which AlphaGo won its first game of Go, a group of museums from the Netherlands saw the birth of the first AI-generated Rembrandt painting, The next Rembrandt, a work that a computer had created from scratch after analyzing all the known paintings of the artist.
Such works are obviously raising new questions about how principles of copyright law, such as authorship, infringement and fair use, will apply to content created or used by AI. The topic is quite complex because, as we have seen above, these programs follow Deep Learning processes and are trained to generate such outputs by partly exposing them to large quantities of existing works, such as writings, photos, music, paintings and other works of art. So how do we deal with sources? And how do you manage the copyright of a work created by a machine?
It is a bit opposite to the case of NFTs in which the artist is perpetually linked to his work through blockchains. The chellenge for works generated by AI is all to be played on the tables of those responsible for managing copyright, such as the U.S. Copyright Office which recently pronounced itself regarding the comic Zarya of the Dawn by Kristina Kashtanova: “As much as the text, selection, coordination and editing of the book are the work of the author, the images illustrating the comic are not covered by the copyright as they are generated by Midjourney technology which is not human work.” And he continues: “Even if a human person provides the images that Midjourney processes, he does not form the image that is generated and is not to be considered the master mind that creates that work.”
In short, most likely with technological evolution the relationship between AI and copyright will become increasingly complex, because inevitably artists will use new software more and machines will get increasingly better at producing creative content. Many hope that in the future the boundary between what is created by man and what is by computer will never disappear and that intelligence will always prevail over the artificial.
