Everything is created by one person's art

When artificial intelligence creates art

A study examines who is responsible for AI art and what role the humanization of AI plays in this

With the help of intelligent algorithms, paintings are created, poems written and pieces of music composed. Whether Artificial Intelligence (AI) is the ingenious creator of art or just another artistic tool depends on how you report on AI art, shows a study from the Massachusetts Institute of Technology and the Max Planck Institute for Human Development .

In October 2018, the work of art was presented in the Christies action house Edmond de Belamie, which was created using an intelligent algorithm, auctioned for $ 432,500. The portrait was created by an artificial intelligence (AI), Christies advertised the auction. The media, too, often spoke of the first work of art that was not created by a person, but independently by a machine. The money did not go to the machine, but to the French artist collective Obvious. It was they who fed an algorithm with images of real man-made paintings, trained the algorithm to create images on its own, then selected a certain image, printed it out, gave it a name and marketed it. The computer experts, on the other hand, who programmed the artificial neural networks and algorithms used, were neither mentioned nor received anything from the sales proceeds.

“A lot of people are involved in AI art: artists and curators as well as programmers. At the same time, there is a tendency, especially in the media, to equip AI with human-like properties. Then you read: The creative AI independently creates ingenious works of art. We wanted to know whether there is a connection between the humanization of artificial intelligence and the question of who gets recognition for AI art, ”says Ziv Epstein, doctoral student at the MIT Media Lab and first author of the study.

Recognition for artists

To this end, the researchers described almost 600 test persons how AI art is created and asked who should receive recognition for the work of art. At the same time, they determined how strongly the participants humanized AI. The answers were very different, but on average it was found that people who humanized artificial intelligence and not just perceived it as a tool also believed that artificial intelligence deserves recognition for AI art and not the people who work on it Development process were involved.

When asked which people deserve the most recognition in the creation process of AI art, it turned out that recognition was initially awarded to the artists who supplied the learning algorithms with data and trained them. Only then were curators named, and then computer experts. And finally, the crowd was named, that is, the mass of internet users who produce the data with which AIs are often trained. Respondents who humanized AI gave computer experts and the crowd more recognition, but artists proportionally less. A similar picture emerges when questions are asked about responsibility, for example when an AI work of art violates copyright law. Here, too, respondents who humanized AIs saw AI have a greater responsibility.

Reporting influences perception

A key finding of the study is that you can actively manipulate whether people humanize Artificial Intelligence by changing the language used to report on AI systems in art. The creation process can be described by explaining that the AI ​​conceives and designs new works of art in a creative way, only with artistic support. Or by saying that an artist designs the work of art and the AI ​​carries out simple commands that the artist tells her. The different descriptions changed the degree of humanization and thus also to whom the participants attributed recognition and responsibility for AI art from the human actors.

“As artificial intelligence penetrates our society more and more, in the future we will have to deal more with who is responsible for what is created with AI. Ultimately, there are people behind every AI. This is particularly relevant when AI fails and causes damage - for example in an accident with an autonomous vehicle. It is therefore important to understand that language influences our view of Artificial Intelligence and that humanizing it leads to problems in assigning responsibility, ”says Iyad Rahwan, Director of the Human and Machine Research Department at the Max Planck Institute for Human Development and co-author of the Study.