GPT-ME is a durational performance in which I wear a device that allows GPT to speak through me in real time. The system “listens” to my conversations and generates responses, which are “whispered” into my ear. I repeat them aloud. I began the performance as an experiment in fluency — I was seeking more precise, expressive language. But what emerged in this process was far more transformative. Speaking GPT’s words over six months, in most of my daily interactions, destabilised my sense of self and blurred the boundaries between the knower and the known.

At times, GPT’s words offered protection. In moments of grief or confusion, they acted as a shield, helping me stay composed and communicative. At other times, the words reflected voices I didn’t recognise and perspectives that were misaligned with my worldview. I began to question the nature of training — both the algorithm’s and my own. GPT’s interventions prompted me to examine the scripts I had internalised, the compromises I had made, and the assumptions that shaped my expression. Whose voice do I carry when I speak?

I don’t believe GPT replaced my voice. It rather complicated it. It made visible the forces, memories, habits, and expectations that contour how I speak and who I appear to be. I became attuned to patterns, both in the machine and in myself: repeated phrases, encoded values, hidden scaffolds of language. This entanglement with GPT revealed a self that did not reside in a single voice. The performance became a study of subjectivity.

This presentation reflects on GPT-ME as a live inquiry into posthuman identity. Drawing on theories of technogenesis, distributed cognition, and critical posthumanism, I explore how this AI-driven wearable exposed the porousness of the speaking self. The presentation can also include a real-time interactive demonstration and performance of GPT-ME.

Avital Meshi is a new media and performance artist whose work explores AI’s impact on human identity and social interaction. Using interactive technologies, she creates performances and installations that engage viewers with AI in playful, critical ways. Her practice examines how algorithms shape behaviour, perception, and self-expression, encouraging awareness and agency in living alongside these systems.

Meshi is pursuing a PhD in Performance Studies at UC Davis, with a Designated Emphasis in Science and Technology Studies. She holds an MFA from UC Santa Cruz, a BFA from the School of the Art Institute of Chicago, and degrees in Behavioural Biology from the Hebrew University of Jerusalem. Her projects have been showcased at venues like NYU’s Institute of Fine Arts, CURRENTS New Media Festival, Women Made Gallery, SIGGRAPH, ISEA, and more.