In October, Christie's auction house will put on sale for the first time works of art created by artificial intelligence. Machine learning is penetrating deeper and deeper into culture, so lawyers are beginning to ask the question: who owns the copyright for works created by algorithms, and who will receive money for the sale?
To at least try to answer this question, you need to figure out exactly how artificial intelligence creates its works. For this, generative adversarial neural networks are used, consisting of two parts. One neural network studies many real works and tries to create its own. The second neural network evaluates the results of the first until it creates something similar to a real picture, photograph, music or other human creation.
Painting "GCHQ" by artist Memo Akten, sold for $ 8000.
Jessica Fjeld, assistant head of the Cyberlaw Clinic law firm on Internet issues, believes that artificial intelligence is just a tool in the hands of an artist.
Based on this, another question arises: which of the people participating in the training of neural networks gets the rights to the results of work? This person, according to Jessica and her colleagues, should be the author of all the elements used at every stage of the creation of AI. This list includes: training kits, training algorithm, ready-made algorithm and, finally, the final product.
With such a wide range of rights, the author can sell not only the work itself, but also the algorithm that created it. Such an offer may be of interest to collectors who will subsequently be able to create other unique creations.
Painting "Electric Fan" by artist Tom White.
Promotional video:
Sometimes AI authors use copyrighted content instead of free learning kits like ImageNet, SoundNet and Google Art, and this approach to development can lead to litigation.
Thus, artists who use artificial intelligence in their work should not have problems as long as they do not violate other people's copyrights. In the future, artificial intelligence may revolutionize art in the same way that the invention of photography changed painting in its day. Gradient Descent curator Kartik Kalyanaraman believes that neural networks can create new forms of art.
Ramis Ganiev