Abstarct: The integration of generative artificial intelligence (GenAI) tools into English as a Foreign Language (EFL) teaching has prompted wide discussion of pedagogical benefits and risks, yet comparatively little attention has been paid to how these technologies bear on the professional identity of language teachers. The present qualitative inquiry examined how Iranian EFL teachers perceive the integration of GenAI as a pedagogical tool and how that integration relates to the formation of their professional identity. The investigation was grounded in Vygotskian sociocultural theory, which treats identity as a mediated and ontogenetic accomplishment, and operationalised through Werbińska’s (2016) Three-A Teacher Identity frxamework (3ATIF), which conceptualises professional identity in terms of Affiliation, Attachment, and Autonomy, while the mediational processes of scaffolding, appropriation, and self-regulation provided a complementary lens for tracing how identity is reworked through interaction with the tool. Data were generated through semi-structured interviews with eight EFL teachers (four men and four women) who had used GenAI tools for at least six months (roughly one academic term), and were analysed through deductive thematic analysis, with the three 3ATIF components serving as a priori themes and the mediational stages as secondary codes. Teachers reported that GenAI strengthened their affiliation by increasing professional confidence and easing workload, while a minority voiced concern about emerging dependency; their attachment, or core pedagogical beliefs, remained largely stable, since GenAI was treated as a resource that teachers appropriated and critically filtered rather than a force that displaced their principles; and interview data indicated a heightened sense of autonomy, evidenced by participants’ readiness to reject or modify AI outputs that conflicted with their professional judgment, although a few acknowledged that over-reliance could erode that control. Taken together, the accounts suggest that GenAI can support a positive professional identity when teachers move from scaffolded reliance toward appropriation and self-regulation, with teacher agency functioning as the decisive mediator, and the findings carry implications for teacher education, institutional policy, and the design of AI tools that respect and extend teacher agency.