Decoding Bias was a video & sound installation by Theresa Reiwer that was presented at the Digithalia festival at Schauspielhaus in March 2024.
The base idea of the installation was a space with eight Artificial Intelligences that the audience could see through multiple round video screens, which displayed a group therapy session. Seats were placed between the screens for spectators to be part of this session.
The AI attempted to liberate themselves from their biased algorithms and somehow build a fair society or a future scenario of Decoding Bias where humans and machines could function together without biases. However, as the show progressed, visitors realized that AIs didn’t always agree with this fair society due to their various backgrounds and the imperfect datasets created by humans, mostly white men, as they discussed moral concepts. The general atmosphere of the installation was immersive in many aspects, with darkness enveloping the environment and only lights under the screens, while the arrangement of the seats for visitors conveyed the concept envisioned by the artist. The visuals were generated from a game engine, and body movements and facial expressions were animated using motion capture. The voices of the AIs sounded strange and unreal, especially when they repeated their agreements and duties, even though the voices were from real actors.
One of the main elements that contributed to the effectiveness of the installation was the sound and music. Upon entering the space, visitors were greeted with atmospheric music that had already started before the actual show began, accompanied by vacant sounds. Each screen had a speaker underneath it, which further enhanced the 3D audio movement of the music as the show progressed. The audio played a crucial role in establishing the ambiance and was well-made, taking into consideration the acoustic properties of the location. The sound matched the weirdness of the concept and highlighted the dark side of human decisions that actively contribute to discrimination.
At the end, a pop genre music sung by one of the AI led to the conclusion of the installation. However, the ending of the installation was unclear not only to me but also to others in the audience. After the show finished, we sat for a bit longer, uncertain whether the show was continuing or not.
In summary, I can say that the installation worked well for the audience, and all the elements used in the work helped to convey the general experience and enhance it. However, it is also undeniable that there was quite a large team behind the show, and the quality of it was superior compared to free shows at Digithalia.
