A multimedia performance that explores artificial intelligence and its limits.
Service:
Art Direction
Video Mapping
Animation
Cooperation:
Fura Del Baus
Barcelona, Spain
Year:
2018
Team:
Sandra Rodríguez, Tim Schnettker,
Xénia Dols, Anna Aguilera Sellas, Kira Teuber
Through the visual and auditorial development of the script, the resulting stage design and the final execution, a unique performance was created. How does a machine interpret morality? This multimedia performance allows the audience to switch between the perspectives of man and machine and thus to understand the thoughts and conclusions of artificial intelligence (AI). The machine as a “black box” is a major contemporary issue that deals with future developments and the integration of machines into society. For this reason, this multimedia performance raises the question of how society will accept the decisions of a machine, becoming part of the conflict between human morality and logical machine thinking.
The Script The visitors are part of a neural network of an algorithm that will get used in the future to support political decision-makers. By maximi- zing rationality, the human error rate in decision-making is to get minimized. The audience is in the training phase, in which the algorithm gets taught morality utilizing questions. They go through different stages, each with two possible answers: right or wrong. To provide the machine with appropriate feedback, they get asked to perform actions via their cell phones. In the course of the play, the AI learns more and more about human morality through the spectators. The piece‘s climax is a classic dilemma: the machine and its logic-driven system commit a fallacy. The visitors got asked whether they trusted the result of a calculator or a reference person more in a calculation task. The answer was clearly in favor of the calculator. This logical argument was used at the end to introduce a false conclusion. If people trust the calculator more than their reference person, they put faith in a machine instead of a human. That is a logically incorrect result ending up in the system degrading the human as a decision-making being. The training is over. It concludes that human morality cannot contribute to logical decision-making; in other words, humans should not make decisions. The piece questions the trust in the logic of artificial intelligence and the resulting outcomes, which may be logically correct in themselves, but disregard moral, human dimensions.
The Stage The installation consists of a walk-in room with walls made of a semi-transparent net called mesh, which gets covered with video projections from two sides. Due to the translucent fabric, the opposite sides get projected, too. That creates an immersi- ve space that is filled with video animations all around. In the centre of the room is a microphone measuring and converting the visitors‘ volume into video signals. The mesh reacts to this volume by becoming spikier and starting to move. It gives the audience direct feedback on their behaviour and allows them to shape the visuals.
The App With the help of software that visitors can install as an app on their cell phones beforehand, the visuals can get controlled remotely with restrictions. That makes it possible to play messages or sounds on the phone or to turn on the flashlight. The app, called “Kalliope”, was developed by Fura Del Baus and made available to the team as part of the performance collaboration program. The show gets directed by two people. One person mixes the live video projections, and the other sends messages and sounds to the visitors. In addition, two large sound systems transport the computer‘s voice and atmospheric sounds.
Through the visual and auditorial development of the script, the resulting stage design and the final execution, a unique performance was created. How does a machine interpret morality? This multimedia performance allows the audience to switch between the perspectives of man and machine and thus to understand the thoughts and conclusions of artificial intelligence (AI). The machine as a “black box” is a major contemporary issue that deals with future developments and the integration of machines into society. For this reason, this multimedia performance raises the question of how society will accept the decisions of a machine, becoming part of the conflict between human morality and logical machine thinking.
Leistung:
Art Direction
Video Mapping
Animation
Kooperation:
Fura Del Baus
Barcelona, Spanien
Jahr:
2018
Team:
Sandra Rodríguez, Tim Schnettker,
Xénia Dols, Anna Aguilera Sellas, Kira Teuber
The Script The visitors are part of a neural network of an algorithm that will get used in the future to support political decision-makers. By maximi- zing rationality, the human error rate in decision-making is to get minimized. The audience is in the training phase, in which the algorithm gets taught morality utilizing questions. They go through different stages, each with two possible answers: right or wrong. To provide the machine with appropriate feedback, they get asked to perform actions via their cell phones. In the course of the play, the AI learns more and more about human morality through the spectators. The piece‘s climax is a classic dilemma: the machine and its logic-driven system commit a fallacy. The visitors got asked whether they trusted the result of a calculator or a reference person more in a calculation task. The answer was clearly in favor of the calculator. This logical argument was used at the end to introduce a false conclusion. If people trust the calculator more than their reference person, they put faith in a machine instead of a human. That is a logically incorrect result ending up in the system degrading the human as a decision-making being. The training is over. It concludes that human morality cannot contribute to logical decision-making; in other words, humans should not make decisions. The piece questions the trust in the logic of artificial intelligence and the resulting outcomes, which may be logically correct in themselves, but disregard moral, human dimensions.
The Stage The installation consists of a walk-in room with walls made of a semi-transparent net called mesh, which gets covered with video projections from two sides. Due to the translucent fabric, the opposite sides get projected, too. That creates an immersi- ve space that is filled with video animations all around. In the centre of the room is a microphone measuring and converting the visitors‘ volume into video signals. The mesh reacts to this volume by becoming spikier and starting to move. It gives the audience direct feedback on their behaviour and allows them to shape the visuals.
The App With the help of software that visitors can install as an app on their cell phones beforehand, the visuals can get controlled remotely with restrictions. That makes it possible to play messages or sounds on the phone or to turn on the flashlight. The app, called “Kalliope”, was developed by Fura Del Baus and made available to the team as part of the performance collaboration program. The show gets directed by two people. One person mixes the live video projections, and the other sends messages and sounds to the visitors. In addition, two large sound systems transport the computer‘s voice and atmospheric sounds.