Book cover: Joanna Zylinska, The Perception Machine: Our Photographic Future between the Eye and AI, MIT Press, 2023.
The Perception Machine: Our Photographic Future between the Eye and AI
Joanna Zylinska · MIT Press · 2023
Why I Read This Book
My current research focuses on how AI-assisted clothing product recommendation functions influence women's self-awareness and sense of autonomy, thus affecting their clothing choices.
I chose this book primarily because the author's focus and explanations resonate with my research and reflections, especially whether AI can accurately identify and recommend relevant clothing products, enabling users to achieve self-understanding. However, what truly captivated me was: what happens when machines participate in how we perceive ourselves?
This question is particularly important to my research on body perception and personalized visual recommendations.
A Moment That Left a Deep Impression
While reading the preface, I encountered a series of images related to machines, perception, vision, and identity. One image posed a simple question:
“Who am I?”
Seeing this image immediately struck a chord with me.
I realized that my body can be observed, measured, categorized, and described. We all want to accurately understand our physical characteristics—and even more so, we want to understand what determines our physical characteristics, what they mean to us, and how we should take control of our bodies.
That moment made me ponder three questions:
“Who am I?
How do I want others to see me?
Who decides what suits me?
I've started to think about whether body image might be one of our gateways to self-understanding.
My current thoughts:
I believe the answer isn't simply that artificial intelligence should stop defining humans.
The relationship between the two is likely more complex.
AI may be better than humans at observing and presenting certain physiological data. However, humans should retain the right to interpret this information and decide its meaning for their own lives.
In other words:
AI can help describe things as they are.
But humans must retain the right to determine its meaning and subsequent development.
This suggests that both AI systems and users may need to change through interaction.
AI needs to set boundaries: What can it measure? What can it infer? Should it never make decisions for users?
Users may also need to develop a new literacy: using machine-generated knowledge about themselves without relinquishing their right to self-definition.
The questions I'm pondering:
Where should the boundary lie between AI measurement and human self-definition?
What rules should AI systems follow to ensure that personalized services support, rather than replace, human autonomy?
To uphold the right to self-definition, what aspects of oneself and AI does an individual need to understand?
If "who I want to be" is unstable, socially influenced, and constantly changing, how should personalized AI systems respond?
Can human-computer interaction become a process of continuous adaptation between the system and the user, rather than the machine simply predicting the user?
Relevance to My Research
These questions are crucial to my research on AI-assisted personalized visual recommendations.
I am particularly interested in a possible shift:
Discovering oneself → Changing oneself → Becoming who you want to be
For women, appearance can be part of this process, but it's not just about looking prettier. Appearance may be closely related to identity, social roles, confidence, self-presentation, and the shifts in a woman's self-perception at different life stages.
Therefore, my question is no longer:
Can AI tell women what suits her?
and more about:
How can AI help women understand themselves and explore who they want to be—without depriving them of their right to choose?
References
Zylinska, J. (2023). The perception machine: Our photographic future between the eye and AI. The MIT Press. https://doi.org/10.7551/mitpress/14471.001.0001.