« Portrait-robot » of Artificial Intelligence
Can a car write a novel? The answer may come as a surprise.
1 The Road, a book about friends travelling between New York and New Orleans, was written in 2018 by a Cadillac.. equipped with a surveillance camera, GPS and microphone connected to an AI, which had been fed with 200 classic Anglo-Saxons. Throughout the road trip, the manuscript was written line by line, on paper. A new kind of founding book or a new feat of artificial intelligence (AI)?
The AI, an artist like any other?
With art, AI has found a new field of exploration. It interferes in all artistic disciplines and pushes back all the boundaries of creation.
Prowess made possible by the progress of learning technologies, machine learning and its evolution, deep learning. Deep learning is a particularly important development since it aims to imitate the human brain by relying on a network of artificial neurons. The more data a machine learning system receives, the more it learns and becomes accurate. And it is the famous Big Data that accelerates the learning curve and allows the automation of data analysis.
The emergence of deep learning has revolutionized the way artists approach AIs who are now able to learn, create with previous data provided to them, create works and paint paintings.
This is the case with Edmond de Bellamy’s Portrait, which has made great impact on the contemporary art market. Sold for $430,000 at auction by Christie’s at the end of 2018 and created by an AI fed by works by renowned painters, this work by the French artists’ collective Obvious beat Andy Warhol on that day, a print of which was sold for $75,000.
The result is often so remarkable that it is impossible to tell with the naked eye which is the result of the AI or the painter’s. This is the case with The Next Rembrandt, launched by MICROSOFT in 2016; the software used lists the characteristics of Rembrandt’s portraits and, coupled with a printer, creates an original work that incorporates all the stylistic codes of the illustrious master of art.
Music is another of AI’s favourite fields. GOOGLE recently allowed its users to create a score in the style of Johann Sebastian Bach, on the composer’s 334th birthday, and HUAWEI completed Schubert’s unfinished symphony No. 8 (1822). The AI of the Luxembourg start-up AIVA completed a piano piece by Dvorak, 115 years after his death, the tool being based on 30,000 scores and the composer’s 115 works. It is through deep learning that AIVA -Artificial Intelligence Virtual Artist – develops the art of composition. The AI analysed 15,000 scores of famous composers using the algorithm to generate a mathematical and intuitive model of music that allows it to produce musical compositions of all styles on demand. Its customers already include PIONEERS, NVIDIA, IBM and GOOGLE, which are all very important because they enable the development of significant cost-effective applications such as music for video games. As a gamer, the most frustrating thing is to find yourself with a very good video game whose atmosphere is damaged by repetitive music when you can play it for several hours.
Thanks to AIVA’s AI, video games will have a substantial music catalogue that will adapt to each situation, making them more immersive and engaging.
In full effervescence, the art market will continue to be transformed under the impetus of the AI, which pushes back the boundaries of creation… but not only. By 2035, the gross value added of the arts, entertainment and recreation sector in developed countries is estimated at $453 billion, and by adding the contribution of AI, at $541 billion, Belgium’s GDP.