Exploring Identity through AI-Generated Selfies: Digital Selfie Creation - Thomas Kassian

 Introduction

I haven't had an assignment that specifically required the use of AI before, so I am excited to explore this. When I think about the implications of AI becoming more advanced, I find it overwhelming. However, I also see this as an opportunity to engage with the technology in a mindful and intentional way. This assignment challenges me to not only experiment with AI but also to reflect on how it shapes self-representation and creativity in digital spaces.

First Attempt

Please generate me an image. A 25 year old man with brown hair and brown eyes who lives in alberta.

Wow. It actually looks a bit like me, even though I put very little detail into my first prompt. I am a very average-looking white guy, but it’s fascinating how AI can generate a face that feels familiar with such minimal input. It makes me wonder how much of our perception of "self" is tied to generic traits and how AI interprets those patterns. This first attempt has me curious - what happens if I provide more specific details? Would it look even more like me, or would it lose some of that uncanny resemblance?

Second Attempt

Please generate me an image. A 25 year old man with brown hair and brown eyes who lives in alberta. He has light freckles, thick hair that almost looks like a muppet, and enjoys filmmaking. He has a moustache and no other facial hair. He also has thick eyebrows.

 

I thought it would be better here. I prompted it to only have a moustache, but it gave me a full beard. The freckles and eyebrows are a bit closer - but it definitely did not get a few aspects of my prompt.

Final Attempt

Generate an AI selfie of me, taking my likeness from the image I sent you. Show me holding a DSLR camera, and a collage behind me. The collage includes sushi, lakes, polaroids of friends and family, volleyball, and tennis.

For my final attempt, I decided to take things a step further by providing an actual image of myself as a reference. I was curious to see how well AI could interpret my likeness when given more concrete data rather than just a text description. The result was surprisingly accurate ... many of my defining features were captured, from my moustache and freckles to my general facial structure. I also said nothing about chips - but it looks like there are Lay's in the corner?

What stood out the most was how AI balanced personalization with stylistic interpretation. While it reflected my general appearance, it still had an AI-generated quality to it, almost like an enhanced version of reality. 

Conclusion

This experiment made me realize how AI can be a powerful tool for self-representation, but it also raises interesting questions about authenticity and digital identity. While the AI-generated images captured many of my defining features, they still have a stylized, almost idealized quality. This makes me wonder about the future of AI-generated imagery not just for personal use, but in fields like media, marketing, and even security.

What’s even more interesting is how AI’s ability to generate images will likely continue evolving. Right now, it still has limitations in precision and nuance, but what happens when these tools become even more accurate? How will we differentiate between real and AI-generated images when they become indistinguishable? There are ethical considerations as well ... who owns an AI-generated version of my face, and could it ever be misused?

Beyond that, this project reinforced the idea that we "don't know what we don't know" when it comes to AI’s potential. We’re still figuring out the best ways to use it, and new applications will continue emerging. It’s both exciting and daunting to think about how AI will shape creativity, self-expression, and even our understanding of identity in the future.