Avital Meshi


Don’t Worry, Be Happy (2020) Is it possible to look at a person’s face and determine how they feel? As Artificial-Intelligence systems continue to develop, tremendous resources are invested in figuring out how to objectively infer people’s feelings based on their facial analysis. AI emotion recognition systems are able to detect faces and return confidence levels across a set of emotions such as anger, disgust, fear, happiness, sadness and surprise. But to what extent can they truly reveal a person’s emotion? Such systems already operate in our environment without us knowing that they are there. However, it is important to remember that AI systems are not ‘things-in-themselves’. These systems are intermingled with people. Our entanglement with them portrays a mutual constitution of agency which further provides a reciprocal ability to act.

In the live performance ‘Don’t Worry, Be Happy’, the artist is strapped to an electric chair. Her face is constantly detected by an emotion recognition AI system. As long as she is detected as ‘Happy’ she is safe. However, if any other emotion is observed, she receives a non-lethal, 2-second long, electric shock to both her arms. During the performance the artist changes her apparent behavior in order to free herself from the ‘punishment’ that the AI system delivers. Yet, under the threat of getting shocked, for how long can she perform this exaggerated facial expression so that the machine continues to ‘read’ her as ‘Happy’?

The kafkaesque apparatus presented in this performance is inspired by posthumanist performativity ideas. The dynamic process of intra-activity between the body and the machine allows meanings and patterns of marks on bodies to be constantly reconfigured. Agency flows between the body and the AI system so that the meaning of ‘emotions’ or ‘feelings’ cease to function as simple labels or words but rather they become an ongoing performance of the world in its differential intelligibility.