Deepfake technology has been deployed to create a new digital flirtation act that explores the social prejudices of artificial intelligence.
The Zizi Show invites the audience to choose a drag artist and a deepfake performance to watch. As digital bodies move across the virtual stage, viewers can switch between different AI identities to expose the methods used to generate them.
The project is the idea of artist Jake Elwes, who created the virtual cabaret by training a neural network on videos of 13 drag artists.
“There are drag kings, queens, biologically female drag queens, drag king and trans queens, drag monsters,” he said. “We also made sure there was diversity in terms of race, gender and sexuality.”
Over time, the neural network iterated and improved its outputs until it could produce realistic deepfakes of the performers.
The cast members range from Moi, “a lip-syncing master with an aesthetic best seen through a pair of sunglasses,” to Oedipussi Rex, “a barbarian barbarian, with acts as incredibly incoherent as the gods themselves.
Elwes also generated a shape-shifting host for the show called Zizi by simultaneous training of the neural network on the images of all artists. He describes the result as a “queering” of the data.
[Read: How do you build a pet-friendly gadget? We asked experts and animal owners]
The show is the latest iteration of Project Zizi, a collection of artwork that explores the intersection of AI and drag.
The project started in 2019 with “Zizi – Queering the Dataset”, which studies how facial recognition algorithms misidentify non-cisgender people due to bias in their training data.
Elwes disrupted these biases by retraining a neural network on 1,000 images of drag artists. This led the system to generate smoother faces.
The Zizi Show shifts the project’s focus to performance to communicate these issues to a different audience.
Ultimately, Elwes believes that if artists collaborate with scientists on AI development, they can reshape the technology to better represent our various societies.
Published March 8, 2021 – 15:51 UTC