Research on generative artificial intelligence in educational settings has grown rapidly; however, the evolving process of teacher professional identity negotiation during sustained interaction with these emerging technologies remains insufficiently theorized. This study aims to propose a comprehensive process model illustrating how teachers renegotiate their professional identity when encountering advanced artificial intelligence systems. Adopting the grounded theory approach by Strauss and Corbin (1998), the present study examined 24 active school teachers in Iran, carefully selected through purposive and theoretical sampling methods. Data gathered from in-depth interviews, reflective journals, and chat logs were systematically analyzed using open, axial, and selective coding in MAXQDA. The findings indicate that "dynamic identity negotiation" is a three-phase process comprising Disruption (epistemic shock), Experimentation (experimental boundary-setting), and Reconfiguration (co-production and selective delegation). This complex process is regulated by intervening conditions such as identity flexibility and professional craft valuation. Ultimately, by positioning artificial intelligence as a non-human epistemic other, this model significantly expands the scope of dialogical identity theory.