Label #01: Gender Shell / 2020
“Gender Shell” is an AR installation to visualize gender classification experience, as one of the three Reflective design and social experiment installations in the research project Labeliing: The Reflective Design of Socialized Gender Classification.
Installation 1: "Gender Shell`
Technology comes from humanity, what it build is a reflection of majority view. In this premise, what "body shell" (should) we be? What will majority of databases thinks about our biological appearance? Facing the reflection of social expectation, Which "gender shell" would our biological sex be linked to? Based on the concept of this work above, participant’s gender will be classified by an unsupervised machine learning algorithm, that shows the participant’s picture together with a mosaic of faces it thinks it’s similar. Showing how, without using human made labels of faces from a dataset, the machine would divide these faces into groups.
Gender stereotype texts were collected from anonymous user online through survey, answerrs will pop up randomly according to onlooker's facial movement. As a metaphor that we are all classified by others online only base on our gender identification.
AWARDS: DIS 2020 Student Design Competition WINNER
Inbox
- Design Process
We collected 61 participants’ answers about what is the first description pop up in their minds when they saw the words ‘Male’ and ‘Female’. We used a questionnaire to collect answers to try to avoid bias.
- Design Idea
Answers are visualized as windows 2007 text style input in machine. When participant’s face being detected by camera, with the random facial expression (eg: blink eyes, smile, raise eye brown, nod, turn left/ right, etc..), different answers will pop up on the screen.
The background used half-transparent glitch texture, to faint the boundary between reality and virtual.
The design of Mosaic mask is inspired by face recognition system when people’s face being detected, the clearity will be evovled from blurry to clear, the mosaic square would gradually become smaller until fully accurate. Also, blurring person’s face feature delivered a message of ‘Your look is not important, in front of gender classification, all human are equal, equal to be labeled.’
Labeling: The Reflective Design of Socialized Gender Classification
DIS' 20 Companion: Companion Publication of the 2020 ACM Designing Interactive Systems ConferenceJuly 2020 Pages 575–579https://doi.org/10.1145/3393914.3395900
- CONCEPT:
Living in the generation of data explosion, how would socialized gender classification reflect to our gender norms? How might it provoke human introspection on both gender boundary and gender bias? In this paper, we present Labeling, a project that has combined a critical design mindset with reflective design that described both generated collection of gender labeling and novel application of algorithmic bias. By creating an immersive context as a reflective medium as our present stage, this project aims to explore introspective contemplation on gender bias with participants, and questioning the existential meaning of gender boundary.
C R E D I T S
CHIA-HSIN CHEN Lead Designer
YIFAN ZHUANG AR Experience Creator
More details in related publication:
[1] DIS' 20 Companion: Companion Publication of the 2020 ACM Designing Interactive Systems ConferenceJuly 2020 Pages 575–579https://doi.org/10.1145/3393914.3395900