User Experience Researcher


We designed a way to manipulate and deceive the Zoom facial recognition algorithm.


Design a way to manipulate and deceive perceptual cognition


Munira Kazi, Rebecca Hodge and Devin Wang 

Methods used

Artefact analysis, Information visualisation


24.11 – 08.12.2022


We started the project by researching methods of disguise to get a sense and scope of the different methods of disguise from which we could later narrow down and apply the learned to our project. We looked at examples from the military, and animal kingdom, and optical illusions.

Part of our research was artefact analysis of materials that are widely available for the people and because we were still searching for an entry point for a topic we decided to analyse facial mask, makeup, sunglasses and dupatta (scarf) as disguising materials. All that process helped us to define criteria for disguising and some initial ideas of application contexts. 

The discussion about all four artefacts allowed us to see where we are drawn and make conclusions on which object we perceived that will have the most potential to be explored. We decided to focus on a surgical face mask for our Artefact analysis exercise.

Material and Aesthetic aspects from analysis of surgical face mask
Interactive and Social aspects from analysis of surgical face mask
Political, Psychological, Ecological and Economical aspects from analysis of surgical face mask
Historical aspect and our interest from analysis of surgical face mask

Despite all the research, we were still hanging on the question of what exactly we wanted to disguise from the face – facial expressions or identity in particular.

For our information physicalisation, we worked with a piece of research about how facial masks change the perception of one’s beauty both to females and males. We considered a few different approaches to making it: using pins and ropes, simple blocks but we ended up using cardboard cutouts colour-coded representing bar charts.

Despite the quite abstract concept of beauty, we could have done better by using facial masks as points to express the information and tie it to our project theme which could have given us a potential interesting insight on the matte and helped us forward in our project.

Creating our data physicalisation artefact
Data physicalisation prototype: perceived difference in beauty with and without face mask cover in both females and males

After a brainstorming session we had, we decided to make a game where we gave five subjects a mask they had to wear with a drawn face that they were not allowed to see and a note telling them what facial expression they should exhibit under the mask. The note and the mask’s facial expression were on purpose disconnected. The audience had to guess their emotion.

Wearing a smiley face surgical mask while displaying different emotion 1. Photo credit: Rebecca Hodge
Wearing a smiley face surgical mask while displaying different emotion 2. Photo credit: Rebecca Hodge
Wearing a smiley face surgical mask while displaying different emotion 3. Photo credit: Rebecca Hodge
Preparing the assets (surgical masks with drawn faces) for the experience
Selected masks with tickets of conflicting the mask emotions
Running the experience “Guess the emotion under the mask!”


On the tests, we observed that people are capable to seek information from the visible areas of the face and either trying to guess with high certainty what is the emotion on the hidden parts of the face based on their accumulated experience or looking at the body language of the mask-wearer and gather the needed information from clues such as head tilt, shoulders, posture, etc.


If you liked it share!