Exploring Identity through AI-Generated Selfies: Digital Selfie Creation - Yi Wang

 

The experience of representing myself through AI was an interesting one, as it meant that I had to represent myself through text prompts. Through this, I had describe myself, using a subjective process that that required naming as many personal features as I could name. This is similar to what Wu et al. (2021) discussed, which is that an AI utilizes the input and feedback that a human provides to create a output you would desire. Just as the output generated by the AI summary is based on the priorities that the human sets, the output generated by the AI image was based on the priorities that I set as well. 

The prompts were centered around physical characteristics, setting, and style, such as hair type, facial characteristics, and the style of Pixar. Evidently, this has significantly influenced the final image. This demonstrates the point that AI does not generate content independently but is rather influenced by the structure and content of the prompt provided. This is linked to the point made in the paper by Chubb et al. (2022) regarding narratives and representation in AI, which still relies on conventional representations rather than individual complexities.

Although there are certainly similarities between the AI generated selfie and my self perception regarding things such as the overall appearance or mood, there was also a certain level of simplification that took place. This is further reinforced by the idea that although AI can create images that resemble us, it is impossible for it to fully resonate with the essence of our identity.

Comments

  1. Dr. Jessica Laccetti7 February 2026 at 11:08

    Thank you for sharing! I appreciate how you connect your experience to Wu et al. and Chubb et al., particularly your insight that AI outputs reflect human priorities and rely on conventional narratives rather than individual complexity. Your observation about simplification and the difficulty of capturing the “essence” of identity is compelling. Building on this by anchoring your ideas in a specific visual detail from the generated image would deepen the analysis even further. I'll look forward to seeing how you develop this in Assignment 3.

    ReplyDelete

Post a Comment