Loading...
Home / Art World / ARTIFICIAL INTELLIGENCE – Interview with Angela Saltarelli of Studio Chiomenti

ARTIFICIAL INTELLIGENCE Collecting an algorithm is possible

Artificial Intelligence, like a great revolution, attracts all the fields of human activities.

Even in the system and in the art market, Artificial Intelligence is finding its space, being increasingly within the reach of artists, gallerists and collectors.

Recently, the sale of artworks created with Artificial Intelligence has achieved incredible results.

The pioneer of the event was the London auction house Christie’s , which on October 23, 2018 auctioned the work “Edmond de Belamy”, created by the collective of Parisian artists “Obvious” thanks to the use of Artificial Intelligence, and more specifically, to an algorithm.

The sale is a record: the work was awarded, in fact, for $ 432,500 with an initial estimate of $ 7,000 to $ 10,000.

Can Artificial Intelligence be considered only a new medium for art?

Actually, the implications are many: its use on platforms for the sale of artworks, in which AI agents learn the preferences of the users so as to show in the future only works that match their tastes; or even the use for false detection, a function that would help in solving a problem that affects most of the art world.

To address in detail the close relationship that there is between Art and Artificial Intelligence, we talked with  Angela Saltarelli, a lawyer from Chiomenti, to find out what developments there will be for the art world, with an eye on the certificate of authenticity …

  • How is Artificial Intelligence defined from a legal point of view?

The doctrine has provided several definitions in the field of artificial intelligence. Essentially, artificial intelligence (often abbreviated with the acronym AI, from the English term “Artificial intelligence“) is the scientific discipline aimed at developing entities that can carry out tasks or problems usually solved by human intelligence such as: learning, understanding and reasoning, but also creative activities.

Mankind has been trying to build intelligent entities over the centuries, but it is especially since the twentieth century that he began to discuss non-human, or artificial, intelligence.

Two types of artificial intelligence are identified: artificial intelligence in the weak sense (Light AI) aimed at imitating only certain aspects of the human brain, and artificial intelligence in a strong sense (Strong AI), aimed at creating computers capable of simulating the human brain almost completely. Artificial intelligence in a strong sense is developing thanks to the use of Machine Learning systems, that is a set of mathematical-computational methods that allow machines to learn automatically through experience. A branch of Machine Learning is, then, Deep Learning, made up of learning models inspired by the structure and functioning of the human mind.

Considering the evolution of robotics and artificial intelligence in the last decades, the European Parliament adopted on February 16, 2017 the “Resolution” containing some recommendations to the Commission concerning the rules of civil law on robotics (the “Resolution”). This Resolution examines in a very general way intellectual property rights matters, including copyright, relevant in the artistic field.

  • Which role is A.I. having or will have in recognizing art fakes?

In the last decade artificial intelligence has been developing also in the artistic field, in some services related to art, for example, in recognizing fake artworks. The technology usually available to art historians – infrared spectroscopy, radiometric dating, chromatography – is going together with artificial intelligence.

In 2004, some researchers at Dartmouth College developed an algorithm capable of assessing whether a painting, a drawing or a print was false by examining the brush strokes and other technical characteristics and comparing them with the artist’s style. This algorithm, for example, by analysing “The Madonna with child by Perugino confirmed its authenticity, indicating also that it was realized by the Umbrian master with the help of four scholars.

Another example in this context is the artificial intelligence system created in 2017 by Rutgers University, which developed, together with the Atelier for Restoration & Research of Paintings in the Netherlands, a neural network educated thanks to a database of about three hundred works by artists such as Picasso, Matisse, Schiele, decomposed in more than eighty thousand strokes. The AI learned the distinctive characteristics of the works, analysing the individual brushstrokes of each painting, evaluating the stroke, the curvature, the speed with which the colour was applied, in order to understand whether the work was false or not. The algorithm was trained by introducing also false works into the database.

The percentage of success in the evaluation of fakes was about 80%, according to the researchers.

In the future it is likely that the success rates in the recognition of fakes by artificial intelligence systems will become even higher. These systems can therefore be used not only to create the expertise on a work, but also in the judicial field when assessing the authenticity of a work when its paternity is the object of a dispute.

  • Which documentation shall accompany the works created with A.I. and what will be the role of the certificate of authenticity with A.I. works?

Artificial intelligence is used not only to recognize fakes, but even to create new artworks. In this regard, we speak of “computational creativity” and we can draw a distinction between two types of works: the works created by man with the aid of artificial intelligence and the works created autonomously by the AI.

For the first category of works, artificial intelligence is only a tool that the artist uses to create them. Therefore, as this work is created by man with the support of artificial intelligence, even this type of works must be accompanied by the certificate of authenticity and any documentation relating to its origin, such as: invoices, purchase documents or donation contracts, exhibition catalogues.

The second type, that is the so-called “computer-generated works”, is instead made up of works created almost independently by artificial intelligence, using Machine Learning algorithms. In this case, man can affect, at most, only on the learning stage of the system by setting operational criteria and providing a database, often photographic, to the machine. The data provided by the programmer are reworked by the machine, creating new works.

Since the certificate of authenticity is a document intended to attest to the authorship of a work, it is first necessary to wonder whether copyright can arise from a work created autonomously by artificial intelligence systems and, if so, who will be the copyright owner. While in some Common Law countries (such as Honk Kong, England, India, etc), copyrights are attributed to the programmer, whereas according to European and Italian legal framework copyright does not seem to arise from “computer generated works”. Even if this legal framework changed, it would not seem, however, that the machine can be considered the owner of patrimonial, or even more, moral rights. Therefore, the certificate of authenticity for “computer-generated works” seems to lose importance as an expression of the moral right of paternity, not clearly attributable to a machine and that will, at most, be issued by the subject who will be recognized in the future author of these types of works.

  • In view of a larger use of the AI. how could the role of the artist be reconsidered?

Artificial intelligence poses important challenges about the possible replacement of many professions, especially those of a manual nature. According to some critics and art historians, the artist’s role could be completely downsized in the next few years, reducing himself to the selection of operational criteria or images supplied to the machine during the learning phase, or even eliminated.

According to others, however, the art created by AI, as already happened at the birth of photography, will not replace other forms of art, but will be a new form of expression, as a new artistic medium of a conceptual nature. According to this approach, human creativity will not be replaced, but must be identified in the codes of programs created by man and that will be used by the machine.

  • The museums from a legal point of view, what measures should implement to accommodate in their collections A.I. works, created independently by the machines?

Some museums began to welcome digital artworks or “computer art” in their collections since the ‘60s : for example,  Victoria and Albert Museum or the Mori Building Digital Art Museum Tokyo, recently inaugurated, have a computer generated art collection. By computer art, we mean all those works created more or less autonomously by the machines.

For this type of art, museums should pay attention to the legal problems resulting from the extreme ease in reproducing these works, as well as the difficulties in conservation and exhibition, considering the rapid obsolescence of technology. Moreover, for the works created autonomously by artificial intelligence, museums shall consider the extremely uncertain legal framework regarding the ownership of copyrights and, therefore, affecting also the rights of reproduction and economic exploitation on such works. Furthermore, it will be necessary to ask what role critics and art historians will have in relation to the works created autonomously by the AI.

Artificial intelligence is used, on the other hand, with fewer problems from a legal point of view, by museums to improve their services and fruition. For example, the Philadelphia Barnes Foundation has recently used a technology developed by Rutgers University instructing it through machine learning, in order to create a sort of artificial art critic, able to interpret and recognize, in a simplified way, works and objects of art. Finally, the Akron Art Museum, Ohio, offers its visitors the opportunity to communicate with “Dot”, a chatbot that offers a digital guide to the museum’s collection. 

What’s left is to discover how Artificial Intelligence will gain the trust of artists, art galleries and collectors from all over the world…

Below is a video content by Andrea Concas  “Artificial Intelligence in the art world” that answers some important questions: how is it used? For what?

Subscribe to our newsletter

By signing up I accept the terms of the privacy statement

Discover the other secrets of the art world...

The Castelli Method
What is an art gallery
Discover the art secrets told by the greatest gallerists. Download the free VOL III eBook!

Art Rights - all rights reserved any reproduction is strictly forbidden®© 2018

%d bloggers like this: