A Night of A.I. and Drag is an evening of live performance, taking place on the Jubilee weekend on Saturday 4 June, in which artists Jake Elwes and Me the Drag Queen utilise Artificial Intelligence and real-life drag to present a unique performance about queerness, cabaret and technology as part of Among the Machines at the Zabludowicz Collection. The artists have invited Charlie Wood, Egon Centrik, Mahatma Khandi, Joy Less, Bourgeoisie, WET MESS and Oedipussi Rex to collaborate, as they probe and foreground what it means to queer the dataset.

This live performance is part of the Zizi Project (2019 – ongoing), a collection of works by media artist Jake Elwes exploring the intersection of Artificial Intelligence (A.I.) and drag performance. Drag challenges gender and explores otherness, while A.I. is often mystified as a concept and tool and is complicit in reproducing social bias. Zizi combines these themes through a deepfake, synthesised drag identity created using machine learning. The project explores what A.I. can teach us about drag and what drag can teach us about A.I.

YASS met Jake Elwes and this is everything you should know.

What is A Night of A.I and Drag and what shall we expect?

‘Zizi and Me, A Night of A.I and Drag’ is an evening of eclectic drag performance; a deepfake drag double-act where we’re using tropes of drag performance, cabaret and musical theatre to challenge narratives surrounding A.I. and society. Expect numbers featuring Me the Drag Queen performing duets next to an AI doppelgänger – a life size deepfake projection of themselves. 

How did you come up with this idea?

The idea came about through collaborating with one of London’s top drag queens – Me the Drag Queen and discussing how my world of A.I and art could collide with her world of live drag cabaret acts. For a few years I’ve been exploring concept of queering artificial intelligence to tackle bias towards normatively in standardised facial recognition datasets, I did this through injecting images of drag kings, drag queens and drag things into the dataset to shift the weights of the neural network into a space of otherness, a queer space. The resulting images from the work ‘Zizi: Queering the Dataset’ (2019) show constantly shifting faces where lipstick, lashes and liner break down the machines understanding of how to recognise a human face in a quite unpredictable and uncontrollable way. The next step was to think about how we can use artificial intelligence as a performance tool, how Me the Drag Queen’s lip-syncs and movement could be incorporated into these cold computational systems.

How can artificial intelligence be combined with real-life drag?

For this show we’re using deepfake technology – or to be more specific a ‘conditional generative adversarial neural network’ – to construct our A.I drag character. The way this works is that we create a dataset of images of our drag performer from many angles, these images then get converted into stick figures (pose detection). The neural network’s job is then to try and create new fake images based on these stick figures, it’s scored on how close its imagined images are to the original images. Once it’s got very good at turning stick figures into bodies, we can then feed in any new movement which will be converted to stick figures and then output a realistic deepfake of the original body. Some of my favourite moments are when this process breaks down though and we see the limitations and deepfake process at work – such as when we input a particular dance move which was never included in the original dataset and results in their wig or gown imploding in a glitchy puff of digital artefacts.

How does drag challenge gender and explore otherness?

Drag is a performance form which is queer, joyful and expansive. It’s not limited to gay men doing female impersonation and recently we’ve seen a rise in drag kings, female and trans drag queens and drag things all blurring the boundaries and challenging notions around gender and exploring otherness. 

What is the biggest obstacle you faced?

We struggled with the project through lockdown having to put any live aspect on hold. Instead, we created a virtual cabaret – The Zizi Show www.zizi.ai – at a time when we’d lost our queer safe spaces and drag venues. Working with 13 of London’s top drag artists we created a deepfake drag cabaret where you can switch between the bodies and numbers performing online mid lip-sync.

What is the message behind this performance?

We’re trying to imagine a queer utopia, reaching towards a future we want to achieve which is hopeful and where humans and machines collaborate rather than a dystopic one with A.I taking over. Although at the same time we are aware that introducing some queerness and otherness into A.I systems is sorely needed at a time when the biases are being hard coded into these systems leading to algorithmic social injustice. We also wanted to think about how we can tackle the exploitative and non-consensual nature of deepfake technology by reclaiming the process and using it in an ethical controlled way and as a celebration. But for now, we hope to engage new audiences in conversations around A.I through drag and enjoy as Zizi performs in all their fabulousness!

More information here:

https://www.zabludowiczcollection.com/events/view/zizi-me-performance-lecture

Event brief:

Zizi and Me, A Night of A.I and Drag
Join us at the gallery this Bank Holiday Saturday for an evening of eclectic drag performance. Expanding on The Zizi Project, Jake Elwes and Me the Drag Queen invite drag artists Charlie Wood, Egon Centrik, Joy Less, Bourgeoisie, and Oedipussi Rex to join them in presenting an array of performances which probe queerness, cabaret and technology.

The Zizi Project (2019 – ongoing), is a collection of works by Jake Elwes which explore the intersection of artificial intelligence (A.I.) and drag performance. Drag challenges gender and explores otherness, while A.I. is often mystified as a concept and tool and is complicit in reproducing social bias. Zizi combines these themes through a deepfake, synthesised drag identity created using machine learning. The project explores what A.I. can teach us about drag and what drag can teach us about A.I.

This event is FREE, booking recommended.