A text box, a description, and a render button. All of these components simplify the functions of a text to image AI machine. Such simple gestures can result in highly controversial imagery for a multitude of reasons. This can go from the underlying hints of racism to even the male gaze.
In our previous presentation for the Vis 149 seminar, we decided to explore the limitations of AI when it came to indirect descriptions for an individual's identity. The most impactful thing that we took away from that project was the manner descriptive words, such as "angelic", impacted the results of an AI text to image machine. This is what we wish to explore further moving forward.