Generative AI is everywhere. With machine learning models becoming more powerful, more customizable, and more accessible (all of the hardware and software I used to train the models for this project is available online to anyone with the necessary programming skills, for free), it’s vital that the people most impacted by these tools understand not only what they are and aren’t good at, but also what they are and aren’t good for.

As a writer and programmer both excited by and skeptical of generative AI tools, I wanted to figure out what value they might hold for me and my writing and help other writers do the same. Would they pick up on aspects of my voice and style? Might seeing them replicate patterns in my work help me understand my tendencies as a writer? How much of *me* goes into my writing, and how much of that voice and identity can a machine learning model imitate?

This website showcases some of the outputs of my models, framed as alternate versions of myself. Each page is meant to emphasize particular trends I noticed in the outputs which reveals something about the models or about my writing. I’ve also given some context for and reactions to its impersonations. If you’re interested in the conclusions I’ve drawn from the process, you can read my dissertation [link coming later this year]. This site, however, is a place for you to explore and artistically engage with my “data doppelgangers” for yourself, so you can draw your own conclusions.