Translation: Jon Alkorta Martiartu, PhD (UPV/EHU)
Art and artificial intelligence
The ever-increasing amount of data that a computer can process is paving the way for unthinkable works to come to life. Artificial intelligence at the service of the arts is making it possible to bring to life creations that never existed but could potentially have existed. We will look at two in particular: the ‘Next Rembrandt’ and Beethoven’s Tenth Symphony.
The Next Rembrandt
Back in 2016, Eva Astorga wrote an article in Mistérica talking about a project that brought together art, big data and AI to build this excellent Dutch painter’s ‘average’ (if we could call it so) work. The process had several statistical phases that determined the fact that Rembrandt used to paint more men than women, his attire, and that he had some facial hair. From there, the faces, tones and brushstrokes of his paintings were analysed to eventually build a 3D canvas that Rembrandt never painted.
The project was sponsored by ING, Microsoft, the Technical University of Delft and the Mauritshuis and Rembrandthuis museums. The new Rembrandt was presented on 5 April 2016 in an exhibition in Amsterdam, the place where Rembrandt lived and worked, and then toured several European capitals. I actually had the opportunity to see it at the Fundación Telefónica in Madrid before the pandemic. And the result is unsettling, a face that you can’t deny retains that air of shadows and depth of Rembrandt’s works but which close-up is a far cry from the original creations. Emmanuel Flores, head of the project seems to agree with me:
When I am asked if a human would have painted it better, I always say yes. Because the ability to learn and ‘make sense’ is something in which modern systems are still beginning to learn. As I said earlier, humans have the ability to invent upon so-called ‘mistakes’. That is a great advantage over a machine. The machine is simply more ‘objective’.
If you want to delve deeper into the creative process of ‘The Next Rembrandt’ and read an interview with Emmanuel Flores, here are two references to articles by Eva Astorga:
The Next Rembrandt: Big Data made art
Interview with Emmanuel Flores, who is currently the head of technology and innovation at the creative international agency J. Walter Thompson – Amsterdam.
Beethoven’s Tenth Symphony
We know that between 1824, when Beethoven composed his greatest work of art, the Ninth Symphony, and 1827, when he died, this giant of music was composing the Tenth Symphony, which could potentially have been his greatest work. But only musical sketches of what this composition might have been have survived.
There was already an attempt in 1988 when musicologist Barry Cooper, who reconstructed it from sketches and fragments, released two recordings, one conducted by Wyn Morris and the other by Walter Weller. Logically, this was a free version by a composer separated by a century from the death of the genius composer.
But now another attempt has been made to bring this work to life using AI: a team of music historians, musicologists, composers and computer scientists have pooled their different skills so that a computer analyses a whole series of variables and takes into account all of Beethoven’s previous musical compositions so that the final work has the feel of the genius’s compositions. The whole project has been promoted by Telekom telephone company, which is based in Beethoven’s hometown of Bonn.
You can listen to a fragment of this experiment here:
Once again, the question arises as to whether this composition has the genius’s feel or just a computational construct… These are truly the times of AI.
Given these two significant examples, I pose the following questions: will the works of the great masters be forged? Will creators use artificial intelligence to improve their results? Are we coming to the end of art?
Museo del Prado’s website
But AI applied to art is not always used to create brand-new works, it is also used to display contextual information. Specifically, Museo del Prado is a pioneer in the way it searches for information about its collections. Here, the AI engine acts by interpreting the user’s queries in a semantic way to provide all the necessary information.
It also offers the creation of a customisable multi-layered timeline in which the work you are searching for can be put in context and placed in a temporal relationship with other masterpieces. The project was carried out by the Prado Museum and the Telefónica Foundation, making the Prado website a benchmark for museums on the Internet.
Museo del Prado, with Telefónica’s support, offers an augmented reading tool that will provide historical context to works and authors. More information in the following link:
Museo del Prado sits at the top regarding AI application to its collections.
Mindfacts: Esguinces mentales
Here are a couple of episodes from the podcast Mindfacts: Esguinces mentales related to the topic at hand, dealing in this case with AI:
AI for beginners – Mindfacts: Esguinces mentales – Podcast en iVoox
AI for experts – Mindfacts: Esguinces mentales – Podcast en iVoox — www.ivoox.com
OpenEdition le sugiere que cite este post de la siguiente manera:
Dr. Pedro Ortega (10 de febrero de 2023). Las artes y la inteligencia artificial. El blog de Tartalo. Recuperado 19 de marzo de 2025 de https://doi.org/10.58079/unve