Due to the phenomenon of data collection only fixes one very specific slice of reality, such as people, places, and contexts, instead of capturing the whole world in all its complexity, that model learns from this one-sided bias. This raises a question that I wonder, who was able to be contribute to the dataset and who is being left out, did participants give informed consent that they can withdraw? Also who labeled the data, how were they trained and paid, and whose norms shaped the instructions for tricky cases.

Provenance gives me a broader view and a clearer logic, and understanding the model’s and dataset’s provenance grounds my creative process. It reminds me to ask who benefits and who bears the burden, and to hold the double edge of technology, and its power to open doors and its tendency to exclude or misrepresent. With that context, I frame ideas more responsibly that I design interactions that play to the data’s strengths, add credit sources and licenses, and make consent and opt-out visible.

Assignment:

https://editor.p5js.org/zw3421/sketches/eUicWY3wO

For this assignment, I designed a game where your nose controls a racket with an emoji “🏑”. I let a single ball bounce around the canvas, and if the ball touches the racket where the nose position is detected, it bounces back and keeps playing. I did not mirror the camera, so it’s a bit hard to control the nose tracking, I might want to change this in the future.

Demo:

https://drive.google.com/file/d/1hZdsVqrKTUkGfOJynwGB926TznHfD8pB/view?usp=sharing