Brief

COMUZI was commissioned by the Mozilla Foundation and curator Irini Papadimitiriou of FutureEverything to produce a speculative provocation responding to the theme of Better Machine Decision Making for MozFest 2019's Arts & Culture Experience.

Client

Mozilla Foundation (for MozFest 2019)

What we did

COMUZI decided to explore the topic of facial recognition.

Facial recognition is currently being framed as a technological innovation powered by machine learning which improves security, becoming the passcode for your phone and pin code for your mobile payment capability.

But it has not being embraced everywhere, governments, polices forces are using facial recognition software to police marginalized communities.

However the people have been responding, hacking their own ways to evade mass surveillance.

There are a number of artists who have explored this area of work and proposed different interventions that people could employ.

Inspired by an academic paper, exploring how infrared light could used to evade facial recognition systems. COMUZI's speculative provocation for MozFest 2019 was the Invisible Mask.

The Invisible Mask is a baseball cap with an array of tiny infrared LEDs wired to the inside of a baseball cap to project dots of light onto the wearer’s face.

These dots are invisible to the human eye, however the lights will confuse the computer vision models in facial recognition systems and make the face unidentifiable.

Results

COMUZI presented a work in progress Invisible Mask at MozFest 2019.

COMUZI also created an immersive exhibit allowing attendees of MozFest 2019 to come and see the Invisible Mask in action, while also learning on how to train a facial recognition model on a mannequin.

We were careful with our presentation of the Invisible Mask as we are aware of the political angle that will be associated with the project.

It is something that we would like to contribute to the discussion on the attack on human agency and autonomy by facial recognition technologies utilized in public security environments.

However, it is important to not create a state of algorithmic anxiety, hence it was important for us to get attendees to learn how to teach a machine learning model.

This project is a work in progress, COMUZI hopes to further develop the Invisible Mask further and allow the project to tour globally to get feedback from a global audience on the topic of facial recognition.

We are grateful to The Mozilla Foundation and Irini Papadimitiriou for allowing COMUZI to produce an artefact that is very provocative.