Experiment 6

After experimenting with Yolov11 and OpenCV, I wanted to convey the idea of personal data used for AI image generation but displaying it in the raw form where real-time image are captured to reconstruct a new image. Despite the existing real-time image generation tool of StreamDIffusion, I wanted to remove the beauty part of AI generation and represent the risk posed with AI generated content with personal data.

Program used: ChatGPT, Python, OpenCV, Yolov11

Method

The process of reconstructing does not involve any AI image generating models but in a less complex method of linear blending.The function for Yolov11 was to identify human and detect human from the webcam and taking screenshots from the live feed to reconstruct the image on the other window. Therefore there will be 2 windows, one being the window for webcam and data loggigng and the other for the reconstruction of the image to replicate the process of diffusion for AI image generation.

Observation

The idea does have flaws as to why we would ever need a cookie prompt shield program if it's mandatory, hence the experiment itself still lacks a substantial evidence to build up the narrative to resonate with the audience. However, I still believe it brings attention to cookie collection prompts, as we come to get comfortable with ignoring such prompts, it loses the awareness behind what cookie collection when people click "Accept" without second thoughts. These also lead to other kinds of data collection that we seem to not care when we accept the terms & conditions without reading it.