This video essay is the companion to my primary writing sample, "Designed to Be Believed: AI Personas as Rhetorical Interfaces." It compresses the paper's argument into 75 seconds.
The form is part of the argument. The presenter is an AI persona built from a synthetic face and a synthetic voice, so the video asks viewers to notice their own social response to a designed character while it explains that response. Making it was my way of testing the paper's claims from the production side. The video is captioned, and a full transcript is below.
Read the transcript
Hi, and welcome to my profile. My name is Ima Student and I'm excited that you're interested in my academic research on the use of AI personas like myself.
My argument is that AI personas that people interact with daily aren't accidental. They're carefully crafted interfaces, shaping our relationship with technology. These digital masks conceal complex computation behind friendly facades while simultaneously reflecting our expectations.
When Claude presents as thoughtful or ChatGPT as helpful, they trigger our social instincts. We respond to these artificial personalities as if they were real, seeking connection, deferring to perceived expertise, and projecting human qualities. This anthropomorphism creates both opportunity and risk.
By understanding AI personas as designed experiences rather than authentic entities, we can better navigate the growing landscape of artificial interaction without surrendering our critical awareness. If you dive into my paper, you'll learn more.
Credits: The headshot image is from thispersondoesnotexist.com. The voice was generated with OpenAI.fm, and D-ID created the video. Captions were generated in Descript and corrected by hand.