This assignment was my first experience using any form of generative AI to create an image. After looking at a few of the sites, I opted to use Artbreeder to generate my selfie. With this, even after adding on new prompts to make the image fit to my own personal traits, AI lacked the ability of fully creating self-representation. No matter how many prompts I used, both specific and not, the image still falls short of the message of self identity I strived to create.
I began this process by listing aspects of my physical appearance, such as, mid length brown hair, brown eyes, British decent, silver jewelry. With these vague details it is almost impossible for AI to create an image that looks exactly like an individual, and even if I had gone into more in depth with facial structure, it would still not output an accurate image. From here in this activity, I decided to add more prompts representing some things I associate with my self-representation. Additional prompts including, explaining I was a university student and that I enjoyed listening to music in my free time. However the AI seemed to ignore these things, not giving me headphones or books, but instead just changing the clothes worn in the image. Therefore I found AI inaccurately follows the prompts provided, and seems more of a generalization than aligning with my actual self-perception. I do not think anyone I know would look at the image and think it was supposed to be me. With this being said, AI is a huge part of our society today, so I found this activity important to criticize its accuracy and representation.
Comments
Post a Comment