(non) adversarial living companions
A design research project examining algorithmic impacts on material culture



(non) adversarial living companions explores what it means to live alongside algorithmic systems - not just through screens, prompts, and digital interfaces, but as physically embodied, attuned presences woven into our daily lives.

As AI starts to gain form and animacy, how will our relationships with the algorithmic world shift from typical, mediated interactions to something more lived, embodied, and mutual — blurring the boundaries between tool and companion, object and subject, or adversary and an ally.




Each object in this series plays on the uncanny and speculates a near-future in which AI’s physicalization doesn’t follow conventional technoutopian typologies, like humanoid robots or sleek devices, but instead emerges through more familiar forms, sensing, interpreting, and gently intervening.    

These “physical AIs,” invite you to reconsider your relationships—with yourself, with objects, and with the wider algorithmic world—not through warnings or conventional utility, but through small, evocative frictions that surface the invisible and spark new imaginaries for coexistence.



Object #1: .io Chair

The .io Chair is indecisive about being idle and not sat on, but also doesn’t want to be sat on for too long. Its only way of interacting with the world and expressing itself is through the means of text printed on a receipt printer.




The .io Chair is embedded with a grid of pressure sensors in its seat and can analyze how someone sits — perhaps they’re leaning back, shifting, or perched on the edge — physical cues that often reveal underlying emotional states.


This data is interpreted by a large language model housed within the chair, and generates a short dialogue with the sitter, printed in real time on a receipt printer. The result is a playful, affective interaction in which the chair actively responds to the sitter’s presence and emotion, proposing a more intimate and nuanced relationship between human and object.




Object #2: Don’t Stare Mirror

The mirror is very vigilant about how long you stare into it, getting you to reconsider your vanity. But at the same time, it encourages positive behaviors, such as smiling.




A Raspberry Pi running a local facial detection AI model is used to analyze and process one’s reflections. When the mirror deems that one has stared for too long, it will project your eyes looking back at you.




Object #3: Perceptive Fan

The fan is aware of those who are constantly attending to their devices, gateways into algorithmic usage and the digital world. When it detects someone, it gently rotates toward them and emits a soft gust of air to subtly prompt environmental awareness.





The fan detects people in space using a computer vision module, making real-time readings on distance, body positioning, and object detection.

This data is sent to an Arduino that triggers the 270 degree movement of the fan on a motor driven rotating plate. Simultaneously, a rack and pinion mechanism pulls back the latex membrane to generate a small burst of air upon release.