The following is the second and last article of this newsletter dedicated to the relationship between Privacy and Copyright on the one hand, and Artificial Intelligence on the other.
If defending privacy may seem like a... no problema, claiming intellectual property ownership of the original works involved in their education may mean forever shutting down any generative artificial intelligence on the market today and excluding any possibility of its being built in the future.
In fact, to make generative AI work, large quantities of data are needed, be they images, manuscripts or other. And if we wanted to legally acquire the rights to all the information necessary to train an AI, billions of investments would be necessary and to date none of the players on the market today have felt the need to take on this problem.
Those who work on generative AI today have no qualms about drawing from the immense digital databases which, outside the control of any institutional guarantee body, proliferate online. And over time, the more power they gain, the more difficult it will be to obtain recognition from them for the intellectual property of the original works.
«Do you want to know how I got all that stuff into my head? With a brain implant. I have given up part of my long-term memory forever. My childhood." From the film “Johnny Mnemonic” by Robert Longo – 1995
Inspired by a novel by the visionary writer William Gibson, the film "Johnny Mnemonic" tells the story of a data courier named Johnny who, hired by a criminal, must transport a large quantity of information stolen from the powerful multinational Pharmakom and crammed in his brain, running from one side of the futuristic and endless city of Newark to the other.
The cyberpunk style setting accompanies a story with dramatic and dark tones set in a place where, to survive the dangers and pitfalls, it is necessary to give up something important, something that is part of oneself. And if it is normal routine for the inhabitants of Newark to replace parts of their bodies with powerful cybernetic implants, lethal weapons that can guarantee their survival in the infamous suburbs of the metropolis, the normal routine for Johnny is to erase the memories of his childhood to free up enough memory to hide precious databases in exchange for money.
If we conceive of the human body as hardware and the mind as software, can we imagine a future where the mind can also be replaced by knowledge that replaces memories and ideas that replace our way of thinking?
OpenAI was founded in 2015 as a non-profit research organization by Elon Musk and others. The deed of incorporation declares a commitment to research «to advance digital intelligence in a way so that all of humanity benefits from it, without being bound by the need to generate a financial return».
The company has declared several times its intention to carry out "research free from financial obligations" and not only that: its researchers would be encouraged to share the results of their work with the whole world in a virtuous circle where winning would be was all of humanity.
Then they arrived Chat GPT,AI capable of communicating by returning information on all human knowledge, and a colossal investment by Microsoft amounting to 10 billion euros which pushed the CEO of OpenAI, Sam Altman, to officially declare: «When the situation became critical, we We realized that our original structure wouldn't work and that we wouldn't be able to raise enough money to achieve our nonprofit mission. This is why we have created a new structure." A for-profit structure.
«If AGI is successfully created», Altman writes again, referring to Artificial General Intelligence capable of understanding or learning any intellectual task like a human being, «this technology could help us elevate humanity by increasing well-being, turbocharging to the global economy and encouraging the discovery of new scientific knowledge that increases the development possibilities of all humanity." And all this, in Sam Altman's intentions, can be possible without any sharing of his discoveries. If you don't believe it, read here.
Is called Stable Diffusion Litigation the website that promotes the cause of some American lawyers against Stability AI, DeviantArt, and Midjourney, platforms for the automatic generation of text-to-image images. The accusation is that of having used the works of millions of artists, all protected by copyright, without any authorization to train its artificial intelligences.
Lawyers point out that if these generative AIs are trained on a large quantity of creative works, what they are able to produce is only their recombination into new images, apparently original but which in fact infringe copyright.
The idea that copyrighted images should not be used in AI training is rapidly gaining ground among artists and is also gaining important positions in institutions.
New York artist Kris Kashtanova has obtained copyright registration in the United States for a graphic novel entitled “Zarya of the Dawn” whose images were generated using the potential of the Midjourney artificial intelligence. But this is a partial success: the US copyright office has in fact established that the images generated by Midjourney in the comic “Zarya of the Dawn” cannot be protected by copyright, while the texts and the arrangement of the elements in the book, yes.
If for Kashtanova the images are a direct expression of her creativity and therefore deserve copyright protection, the US office instead believes that the images created by the Midjourney generative artificial intelligence system represent a "third" contribution, placing emphasis on the “quantity” of human creativity involved in the creation of the work. In other words, the technological contribution of generative AI can be assimilated to instructions given to another artist who, working on commission, returns content to the author over which he has no control.
Midjourney and all its competitors are based on the Stable Diffusion algorithm and the latter belongs to a category of generative AI systems trained through the use of billions of images which, when shuffled, generate others of the same type. According to Stable Diffusion Litigation, this AI is “…a parasite that, if allowed to proliferate, will cause irreparable harm to artists, now and in the future.”
The images that this algorithm is able to generate may or may not outwardly resemble the images with which it was trained. However, they are derived from copies of the training images and are in direct competition with them on the market. Add to this the ability of Stable Diffusion to flood the market with an essentially unlimited number of images that, in the opinion of lawyers, infringe copyright, we are in for dark times characterized by a completely drugged art market where graphic artists of the whole world will soon end up broke.
In this problematic relationship between human and artificial creativity, technological evolution is proving to be so fast as to make any regulatory adjustment obsolete from its first application.
It seems difficult to imagine that all the players already competing to conquer market shares with their own technologies could be forced to suddenly give up using the databases that have already been available to them for years and on which, in the case of OpenAI, they have invested and they will invest rivers of money.
But if copyright were to be imposed also on the data used in AI training, it seems easy to think that company CEOs will find "a new structure" in which to bring together their projects that guarantees them the freedom of movement they deserve. they need. Perhaps simply by moving their registered offices to places on the planet where copyright has no recognition.
Article of Gianfranco Fedele
Developing fine motor skills through coloring prepares children for more complex skills like writing. To color…
The naval sector is a true global economic power, which has navigated towards a 150 billion market...
Last Monday, the Financial Times announced a deal with OpenAI. FT licenses its world-class journalism…
Millions of people pay for streaming services, paying monthly subscription fees. It is common opinion that you…
