Novice and Experienced EFL Teachers' Perception of the Role of AI in Their Professional Identity

1. Introduction

The incorporation of Artificial Intelligence (AI) into educational environments has accelerated rapidly in recent years, fundamentally restructuring teaching and learning paradigms across various disciplines, including language education (Guo & Wang, 2024; Rudolph et al., 2023). AI-mediated instruction has not only revolutionized pedagogical methodologies but also compelled educators to critically reassess their roles, responsibilities, and identities within increasingly digitized landscapes (Derakhshan & Ghiasvand, 2024; Zhou & Hou, 2024). As AI technologies continue to permeate classroom settings, they introduce a complex matrix of opportunities and challenges that directly influence teachers' professional identity (TPI). This influence is particularly pronounced in contexts where cultural, ideological, and sociopolitical factors deeply intersect with language education (Lan, 2024; Zhao et al., 2022).

TPI is a multifaceted, fluid construct molded by educators' interactions with their sociocultural and institutional milieus, alongside their evolving beliefs, emotions, and instructional practices (Beauchamp & Thomas, 2009; De Costa & Norton, 2017; Li, 2025; Mohammad Hosseinpur et al., 2025). It encapsulates how teachers perceive themselves and are perceived by others within their professional capacities (Villegas et al., 2020). The process of constructing and reconstructing TPI is inherently dynamic, frequently characterized by tensions stemming from conflicting sub-identities or external systemic pressures (Kayi-Aydar, 2019; Yuan, 2019). Such tensions can function as a "double-edged sword," either catalyzing professional growth or precipitating attrition, contingent upon how educators navigate these challenges (Eslamdoost et al., 2020). In centralized and ideologically driven educational systems such as Iran, where foreign language policies are inextricably linked to historical, political, and religious discourses, these tensions are significantly magnified (Rashidi & Hosseini, 2019). For example, English language teaching (ELT) in Iran is frequently framed through an anti-Western paradigm, prioritizing localization over globalization and positioning English as a culturally complex phenomenon (Davari & Aghagolzadeh, 2015; Rassouli & Osam, 2019). Consequently, the adoption of AI, often developed in Western contexts and embedded with foreign cultural or ideological algorithms, is viewed with heightened suspicion. Iranian educators must navigate the unique tension between leveraging global technological advancements and adhering to local ideological mandates that prioritize cultural localization and resist perceived Western digital hegemony.

Against this sociopolitical backdrop, the advent of AI introduces novel dimensions to the negotiation of TPI among Iranian EFL teachers. On one hand, AI provides innovative mechanisms for personalizing learning, delivering instantaneous feedback, and augmenting communicative competencies (Fryer et al., 2020; Guo et al., 2022). On the other hand, it challenges educators to adapt their traditional methodologies and roles, necessitating a transition from primary knowledge providers to facilitators of learning (Chounta et al., 2022). This paradigm shift requires a critical examination of how novice and experienced teachers perceive AI's role in shaping their professional identities. While experienced teachers may leverage their established practices and resilience to assimilate AI into their teaching, novice teachers, who are still in the formative stages of identity development, may encounter more pronounced fluctuations and uncertainties (Fairley, 2020; Pennington & Richards, 2016; Tschannen-Moran & Hoy, 2007).

Despite the expanding corpus of research on AI in education, there remains a distinct scarcity of studies exploring its impact on teacher identity, particularly within the Iranian EFL context (Lan, 2024; Eslamdoost et al., 2020). Existing literature underscores the necessity to investigate how AI-induced transformations affect teachers' perceptions of their roles, beliefs, and practices, as well as how these changes fluctuate across different career stages (Beijaard et al., 2004; Mohammed & Nell Watson, 2019). This study seeks to bridge this gap by examining the perceptions of novice and experienced Iranian EFL teachers regarding the role of AI in their professional identity construction. 

To capture the multifaceted nature of these perceptions, this study employs a phenomenographic approach. Phenomenography is uniquely suited for this investigation as it focuses on mapping the qualitatively different ways individuals experience, conceptualize, and make sense of a specific phenomenon (Hajar, 2021). Rather than seeking a single objective reality, this approach justifies its selection by allowing researchers to capture the variation in teachers' meaning-making processes regarding AI, ultimately constructing an "outcome space" that reveals the diverse ways TPI is negotiated in response to AI integration. Specifically, it aims to elucidate the subtle and intricate ways these two distinct groups, possessing differing levels of experience and identity formation, navigate the opportunities and challenges presented by AI within the unique socio-cultural and political landscape of Iranian EFL education.

2. Literature Review

The infusion of AI into educational frameworks has revolutionized teaching methodologies, assessment techniques, and professional development for educators globally. Within the realm of EFL instruction, AI-powered applications such as intelligent tutoring systems, automated writing evaluators, and adaptive learning platforms have fundamentally reshaped pedagogical practices. Nevertheless, despite the escalating prevalence of AI in educational settings, teachers' perceptions of AI’s influence on their professional identity, particularly in EFL contexts such as Iran, remain critically underexplored.

2.1 Language Teacher Identity (LTI)

Research into teacher identity has proliferated over the past two decades, yielding critical insights into how educators' cognition, instructional practices, professional development, and engagement are molded by their evolving sense of self (Ho, 2023; Nazari & Seyri, 2021; Yazan, 2018). More specifically, language teacher identity (LTI) is broadly conceptualized as teachers' self-perceptions, encompassing their beliefs regarding their instructional roles, pedagogical responsibilities, and ongoing professional evolution (Varghese et al., 2016). This construct is influenced by a confluence of factors, including historical background, sociocultural contexts, participation in communities of practice, and pedagogical experiences (Pennington & Richards, 2016; Yazan, 2018). Additionally, language-related factors, such as native-speaker status and linguistic competence, play a pivotal role in shaping LTI, particularly in EFL settings (Pennington & Richards, 2016). While the outdated notion of native-speaker superiority has been widely debunked, language proficiency remains a key determinant in how non-native teachers negotiate their professional identities, enabling them to transcend marginalization and assert their legitimacy within EFL contexts (Pennington & Richards, 2016).

Given its inherent complexity, LTI defies a singular definition, as it is shaped by an interplay of emotional, cognitive, and contextual dimensions (Beauchamp & Thomas, 2009; Yazan, 2018). Scholars emphasize that identity is central to understanding teachers' emotions, agency, self-perceptions, and professional trajectories, as well as their responses to institutional and sociocultural dynamics (Beauchamp & Thomas, 2009; Beijaard et al., 2023). Farrell (2011) posits that language teachers continuously (re)construct their professional identities throughout their careers, reflecting the fluid and adaptive nature of LTI. Aligning with this perspective, Yazan (2018) defines teacher identity as a dynamic process through which teachers perceive and imagine themselves. This identity is not static; rather, it evolves as educators engage with different communities, collaborate with peers, and act as both active participants and subjects of social perception. Such conceptualizations underscore the premise that LTI is an evolving construct shaped by ongoing negotiation within professional and social environments (De Costa & Norton, 2017; Nazari et al., 2023; Yazan & Lindahl, 2020).

2.2 AI and the Reconstruction of Teacher Identity

The introduction of AI into education has prompted significant shifts in teacher identity construction, requiring educators to adapt to novel roles and expectations (Zhang & Hwang, 2023). As digital technologies redefine classroom dynamics, teachers must reconcile their established professional identities with emerging AI-driven practices (Ho, 2023). This process frequently leads to identity reconfiguration, wherein educators reassess their instructional methods, self-efficacy, and professional agency (Goto, 2021). For instance, AI tools such as intelligent tutoring systems and automated assessment platforms compel teachers to transition from knowledge transmitters to facilitators of AI-augmented learning (Flores & Swennen, 2020). Such transformations necessitate a reconceptualization of professional identity in the digital age (Robson, 2018).

Recent scholarship has further expanded the understanding of LTI by examining its intersection with technological advancements (Barkhuizen, 2021; Varghese et al., 2016). LTI is increasingly recognized as a construct shaped not only by traditional pedagogical interactions but also by engagement with digital tools, including AI (Lan, 2024). In EFL contexts, teachers must navigate dual challenges: mastering linguistic content and pedagogical strategies while adapting to rapidly evolving classroom technologies (Chounta et al., 2022). This dynamic is particularly pronounced when comparing novice and experienced teachers. Novice educators, who are still forming their professional identities, may perceive AI as a supportive scaffold that aids in lesson planning, feedback provision, and skill development (Huang et al., 2021; Lan, 2024). In contrast, experienced teachers, whose identities are more firmly established, may view AI as a disruptive force that challenges their traditional methodologies and professional autonomy (Fairley, 2020).

The integration of AI into language education has also prompted a reevaluation of LTI in relation to pedagogical tools. Liu and Fan (2025) argue that AI-mediated teaching environments can lead to significant shifts in how teachers perceive their roles, reinforcing Zhao et al.’s (2022) assertion that teacher identity is deeply intertwined with instructional practices and the tools employed in classrooms. Consequently, understanding LTI in the AI era requires an examination of how technological integration influences teachers' self-perceptions, professional confidence, and engagement with learners (Zhou & Hou, 2024).

However, the literature on AI’s specific impact on teacher identity remains limited. Emerging studies suggest that AI technologies influence educators' professional self-perception in multiple ways. Lan (2024) found that Chinese teachers underwent identity reassessment as they integrated AI into their classrooms, while Zhao et al. (2022) observed that AI tools restructured teacher-student interactions and role perceptions. Additionally, AI has been shown to mitigate identity tensions by harmonizing multiple professional roles (Alasadi & Baiz, 2023). An ecological perspective further highlights AI’s impact across micro (pedagogical), meso (institutional), and macro (societal) levels, as demonstrated in a study on Iranian EFL teachers (Satvati et al., 2023). Similarly, Tsou et al. (2024) reported that Taiwanese English teachers experienced identity empowerment through collaboration with ChatGPT, transitioning from traditional content instructors to English-medium instruction facilitators. These findings underscore the transformative potential of AI in reshaping teacher identities, though research in this domain remains nascent, particularly in EFL contexts.

2.3 AI in Iranian EFL Education

The integration of AI into Iranian EFL education is in its nascent stages, shaped by a unique intersection of infrastructural limitations and ideological policies. While global EFL contexts rapidly adopt intelligent tutoring systems and automated writing evaluators, Iranian educators often face restricted access to premium AI tools due to international sanctions and internet filtering (Mohammad Hosseinpur, 2026). Furthermore, the ideological emphasis on localization means that AI tools are often evaluated not just on pedagogical efficacy, but on their alignment with domestic cultural and religious values. Consequently, AI adoption in Iran is not merely a technological transition but a sociopolitical negotiation, where teachers must reconcile global digital trends with localized educational mandates, adding a distinct layer of complexity to their professional identity construction.

Therefore, the current literature indicates that LTI is a dynamic, context-dependent construct shaped by personal, sociocultural, institutional, and increasingly, technological factors. The growing presence of AI in education introduces new dimensions to LTI, necessitating further exploration of how teachers at different career stages perceive and adapt to these changes. Specifically, a critical gap exists in understanding how the intersection of career stage (novice vs. experienced) and centralized ideological constraints shapes the phenomenographic variation in AI perception. Existing literature often treats teachers as a monolithic group or focuses solely on technological proficiency, neglecting the nuanced, emotionally charged identity negotiations that occur when AI challenges traditional epistemological roles in restricted sociopolitical contexts. This study seeks to contribute to this discourse by examining how novice and experienced EFL teachers interpret the role of AI in shaping their professional identities. As a result, the present study is an attempt to provide an answer to the following research question:

How do novice and experienced Iranian EFL teachers perceive the role of AI in shaping their professional identity?

3. Methodology

3.1. Design of the Study

This research adopted a phenomenographic design to investigate how novice and experienced Iranian EFL teachers perceive the role of artificial intelligence in shaping their professional identity. Phenomenography is a qualitative research approach that focuses on identifying variations in how individuals experience, understand, and conceptualize a particular phenomenon (Hajar, 2021). This methodology was chosen because it allows for an in-depth exploration of a relatively underexplored and complex topic: the intersection of AI technologies and teacher identity. Unlike other qualitative approaches that aim to uncover universal truths, phenomenography emphasizes the diversity of perceptions and meaning-making processes (Pilarska, 2021), making it particularly well-suited for examining the ways in which AI influences teachers’ professional roles and identities.

The decision to employ this design stemmed from two key considerations. First, the integration of AI into education is a recent and rapidly evolving phenomenon, with limited prior research on its impact on teacher identity. Second, teachers’ perceptions of AI are likely to vary significantly based on factors such as their level of experience, familiarity with technology, and teaching context. By focusing on these variations, this study aimed to construct a comprehensive outcome space, a structured representation of the different ways participants perceive AI’s role in their professional identity. In phenomenography, variation analysis is handled by identifying the critical aspects of the phenomenon that participants focus on. By comparing and contrasting participants' transcripts, the researcher delineates the boundaries between different meanings, organizing them into a hierarchical "outcome space." This outcome space comprises "categories of description," which represent the distinct, logically related ways teachers conceptualize AI's role in their TPI, thereby mapping the structural variation in their collective experiences (Marton & Booth, 1997; Pilarska, 2021).

To ensure methodological rigor and address the interpretive nature of phenomenography, the primary researcher engaged in continuous reflexivity throughout the study. Given that the researcher is an EFL practitioner familiar with the Iranian educational context, there was an inherent risk of projecting personal biases regarding AI integration onto the participants' narratives. To mitigate this, bracketing was employed prior to data collection. The researcher maintained a reflexive journal to document preconceptions, assumptions, and emotional reactions to AI in education, consciously setting these aside during data analysis to allow the participants' voices to emerge authentically. Furthermore, peer debriefing was conducted with two colleagues outside the immediate research team to challenge emerging interpretations and prevent confirmation bias, thereby enhancing the trustworthiness and rigor of the phenomenographic outcome space.

3.2. Participants

The study was conducted in the Iranian EFL context, where the professional identities of teachers are influenced by an intricate combination of religious, cultural, linguistic, and educational elements, as well as limitations imposed by institutional, pedagogical, and socio-cultural factors. To ensure a diverse yet focused sample, 20 EFL teachers were recruited for the study, equally divided into two groups: 10 novice teachers (with less than five years of teaching experience) and 10 experienced teachers (with more than fifteen years of teaching experience). Purposeful sampling was employed to select participants who met specific inclusion criteria: self-reported familiarity with AI tools and active teaching experience. The age range of participants was 24–49 years old, with a gender distribution of 13 females and 7 males. All participants held advanced degrees; 12 held Master’s degrees in Applied Linguistics or related fields, while 8 held PhDs, and were fluent Persian speakers of English.

Participants were recruited via Telegram, a popular messaging platform in Iran, and provided informed consent prior to participation. Ethical considerations were prioritized throughout the study, ensuring confidentiality, privacy, and anonymity. For instance, pseudonyms were used in all reports, and participants were assured that their contributions would remain confidential. Additionally, they were given the option to withdraw from the study at any time without penalty.

3.3. Instruments  

3.3.1. Collaborative Reflection Sessions  

To capture the dynamic and interactive nature of teachers’ perceptions, collaborative reflection sessions were conducted to serve as a vital tool for triangulating individual interview data. These sessions were designed to facilitate group discussions where participants could share, negotiate, and refine their understanding of AI’s role in their professional identity. Rooted in sociocultural theory (Vygotsky, 2018), collaborative reflection recognizes that knowledge development occurs through social interaction and dialogue. Two face-to-face sessions were organized, one for novice teachers and another for experienced teachers, each lasting approximately 75 minutes. A set of guiding questions and prompts was developed to encourage meaningful exchanges (see Appendix A for the complete list of sample questions for both collaborative reflection sessions and semi-structured interviews). These prompts were intentionally open-ended to allow participants to explore their thoughts freely. During the sessions, the researcher of the study acted as facilitator, encouraging quieter participants to contribute and ensuring that discussions remained focused on the research objectives. Audio recordings were made unobtrusively to preserve the authenticity of interactions.  

3.3.2. Semi-Structured Interviews 

In addition to collaborative reflection, semi-structured interviews were conducted to gather detailed, individual insights into participants’ perceptions of AI. Unlike group settings, interviews provided a private space for participants to express personal thoughts and reflections that might not emerge during collaborative discussions. Each participant engaged in a one-on-one interview lasting approximately 15–20 minutes. The interview guide included open-ended questions aligned with the research objectives (see Appendix A). Follow-up probes were used to clarify or expand upon participants’ responses, ensuring depth and clarity in the data collected. For example, when a participant mentioned feeling overwhelmed by technological integration, the interviewer asked, "Could you elaborate on what aspects of AI integration feel most challenging?" Interviews were conducted in English, recorded, and later transcribed verbatim for analysis. This dual-method approach ensured triangulation of data sources, enhancing the reliability and richness of the findings.

3.4. Data Collection Procedure  

Data collection unfolded in two primary stages: collaborative reflection sessions followed by semi-structured interviews. Before initiating the main phase, the instruments, a set of guiding prompts for collaborative reflection and an interview protocol, were validated by three experts in applied linguistics and educational technology. Based on feedback, redundant prompts were removed, and additional clarifying questions were added to enhance the instruments' effectiveness.

A pilot test was conducted with five EFL teachers who did not participate in the main study. This pre-testing stage helped identify potential issues with the tools and refine the procedures for data collection. Following this preparatory phase, willing EFL teachers meeting the inclusion criteria were recruited via Telegram. They signed informed consent forms at the outset, and ethical considerations were explained to ensure their comfort and trust.

The first stage involved collaborative reflection sessions, held in a public high school in Tehran. Novice and experienced teachers attended separate sessions to ensure focused discussions. Each session lasted approximately 75 minutes, during which participants collaboratively shared their experiences, negotiated ideas, and reflected on how AI shapes their roles, responsibilities, and relationships in the classroom.

The second stage consisted of semi-structured interviews, scheduled within two weeks of the collaborative reflection sessions. Participants were encouraged to elaborate on themes discussed during the group sessions or introduce new insights. After completing both stages, participants submitted reflective notes via email or Telegram if they wished to add further thoughts. As a gesture of gratitude for their participation, small gifts were distributed among the teachers.

3.5. Data Analysis  

Data analysis was conducted using phenomenographic thematic analysis, integrating Braun and Clarke’s (2006) framework with Marton and Booth’s (1997) approach to identifying categories of description. The process involved six systematic steps. First, the researchers familiarized themselves with the transcripts from collaborative reflection sessions and semi-structured interviews, taking detailed notes to identify salient statements and recurring patterns. Reflexivity was maintained throughout this stage by consulting the reflexive journal to consciously bracket preconceptions about AI and teacher identity.

Second, relevant statements were compiled into "pools of meaning." For example, a novice teacher’s statement, "AI is like a safety net when I feel uncertain," was initially coded as "emotional support." Through constant comparison, this code was grouped with similar statements about "feeling prepared" to form the broader category of description: "AI as a Source of Confidence." Conversely, an experienced teacher’s remark, "AI undermines my authority," was coded as "loss of control," which merged with other statements to form "Threat to Professional Autonomy." These initial codes were then reviewed, combined, and refined to form overarching categories of description. For instance, codes such as “AI as a confidence booster” and “Fear of obsolescence” emerged from the data provided by novice teachers. These categories were iteratively compared to delineate the boundaries between them, ensuring logical and hierarchical relationships within the outcome space.

Third, the categories were cross-checked against the raw data to ensure coherence and relevance. Redundant or overlapping categories were consolidated to create a comprehensive outcome space that accurately represented the structural relationships between the different ways participants perceive AI’s role in their professional identity. Each category was then defined and labeled to reflect its essence. For instance, the category “AI as a Double-Edged Sword” encapsulated the dual nature of AI as both a supportive tool and a source of anxiety for teachers.

The final report integrated findings from both collaborative reflection sessions and semi-structured interviews, presenting categories alongside supporting extracts. Shared categories identified across novice and experienced teachers, such as “Impact on Professional Identity” and “Need for Professional Development on AI,” were emphasized to highlight commonalities in their experiences.

3.6. Trustworthiness

To ensure the rigor and credibility of the findings, multiple strategies were implemented. Member-checking was conducted by sharing preliminary themes with participants to verify their accuracy and resonance with their experiences. Triangulation of data sources, including collaborative reflection and interviews, further strengthened the credibility of the findings.

An external auditor reviewed the entire data analysis process to ensure objectivity and transparency. The researcher also maintained reflexivity by documenting his biases and assumptions throughout the study, utilizing the aforementioned reflexive journal and peer debriefing sessions. Detailed descriptions of the participants, setting, and procedures were provided to enable readers to assess the applicability of the findings to other contexts, enhancing transferability.

Inter-coder reliability was established through independent coding by two researchers, achieving a Cohen’s Kappa coefficient of r = 0.83. Disagreements were resolved through discussion and consensus, ensuring dependability. By adhering to these principles, the study ensured high-quality, trustworthy findings that contribute meaningfully to the understanding of AI’s role in EFL teachers’ professional identity.

4. Results

To explore the perceptions of novice and experienced Iranian EFL teachers regarding the role of AI in their professional identity, the data collected through collaborative reflection panels and semi-structured interviews were thematically analyzed. Rather than treating novice and experienced teachers' perceptions in isolation, the findings are integrated into comparative thematic categories to illustrate the phenomenographic variation and highlight how career stage mediates the experience of AI integration. Table 1 provides a summary of these categories.

Table 1 

Summary of Phenomenographic Categories of Description: Novice vs. Experienced Teachers

4.1 AI as a Scaffold vs. Threat to Autonomy

A primary variation in the outcome space relates to whether AI is perceived as an identity-affirming scaffold or an identity-threatening disruptor. For novice teachers, AI emerged as a pivotal source of confidence and support, enabling them to navigate the challenges of early-career teaching. NT5 remarked, “AI tools have been like a safety net for me. They provide instant feedback and resources when I feel uncertain about my decisions.” This perception reflects how AI mitigates the insecurities often associated with inexperience. Some novices viewed AI as a gateway to exploration, with NT8 stating, “AI opens doors to new possibilities. It encourages me to experiment with different teaching methods and see what works best for my students.”

In sharp contrast, experienced educators exhibited skepticism and resistance, often viewing AI as a direct threat to their professional expertise and autonomy. ET3 remarked, “I’ve been teaching for years without AI, and I don’t see why I should suddenly change my approach now.” This resistance stems from deeply ingrained pedagogical philosophies. ET5 stated, “If AI takes over key responsibilities like grading or lesson planning, what’s left for us? It undermines our authority as educators.” This theme reflects fears of deprofessionalization, wherein teachers risk being reduced to facilitators rather than creators of knowledge. ET10 added, “I value my independence in designing lessons and assessing students. Handing these tasks over to AI feels like losing control.” To further illustrate this divide, NT11 noted, "AI gives me the scaffolding I lacked in my first year, allowing me to focus on student interactions rather than just survival," while ET11 countered, "My expertise was built on decades of human intuition; relying on an algorithm feels like a betrayal of my professional journey."

4.2 Redefining Professional Identity and Relational Dynamics

Both novice and experienced teachers agreed that AI significantly impacts professional identity, forcing a redefinition of their roles, albeit with different emotional valences. NT4 observed, “AI forces you to redefine who you are as a teacher. Are you still the expert, or are you a guide helping students navigate AI?” Similarly, ET7 reflected, “Our identity as educators is shifting. We’re no longer just instructors; we’re curators of digital experiences.”

However, both groups expressed profound concerns about AI’s impact on teacher-student relationships, emphasizing the irreplaceable value of interpersonal dynamics. NT9 noted, “If AI handles most interactions, there’s less opportunity for meaningful human connection.” ET10 echoed this sentiment, saying, “Building rapport is central to teaching. I worry that AI might dilute that aspect.” Despite these concerns, novice teachers frequently leveraged AI as a motivational tool to foster student engagement. NT6 explained, “Students respond well to interactive AI applications. They find them engaging and fun, which makes learning more enjoyable.” NT4 added, “AI helps me create personalized activities that resonate with individual students.” Highlighting the relational stakes, NT12 observed, "I am no longer just a knowledge dispenser; I am a facilitator who guides students on how to ethically use AI," whereas ET12 reflected, "The rapport I built over 20 years cannot be replicated by a chatbot, and I fear the human element of teaching is being marginalized."

4.3 Navigating Technological Integration: Engagement vs. Overwhelm

The integration of AI presented distinct practical challenges based on experience level. Novice teachers embraced AI to revitalize classroom dynamics, with NT13 sharing, "AI gamifies the learning process, keeping my students engaged in ways traditional methods sometimes fail to achieve."

Conversely, many experienced teachers reported feeling overwhelmed by the rapid pace of technological integration. ET2 admitted, “The sheer volume of AI tools available is daunting. It’s hard to know where to start or which ones are worth investing time in.” ET7 elaborated, “Even when I try to incorporate AI, I often encounter glitches or confusion. It’s frustrating and time-consuming.” ET13 expressed similar fatigue, stating, "The constant updates to AI platforms are exhausting; I spend more time learning the software than actually teaching." Despite this overwhelm, some experienced teachers demonstrated a pragmatic approach, advocating for a balanced integration of AI. ET4 noted, “Teaching is evolving, and we can’t ignore that. But we also shouldn’t abandon proven methods just because something new comes along.” ET6 added, “I try to blend AI with traditional techniques, ensuring that technology enhances rather than replaces core teaching principles.” Furthermore, experienced teachers emphasized the importance of critically evaluating AI tools. ET9 stated, “Not all AI applications are created equal. Some add real value, while others are gimmicky or impractical.” ET1 elaborated, “I test each tool rigorously to ensure it aligns with my goals and meets my students’ needs. Only then do I consider adopting it.”

4.4 Systemic Vulnerabilities: Dependency, Obsolescence, and Resource Disparities

Beyond pedagogical integration, both groups grappled with systemic vulnerabilities, though their specific anxieties differed. Novice teachers expressed existential concerns about their relevance and over-reliance on technology. NT7 articulated this fear, stating, “I worry that one day, AI might replace teachers entirely. If machines can do everything we do, what will be left for us?” NT9 stated, “Sometimes I feel like I’m too dependent on AI. Without it, I wonder if I could manage as effectively.” NT14 confessed, "If the server goes down, I feel paralyzed; I worry I am losing my own pedagogical instincts."

Experienced teachers, while less fearful of obsolescence, maintained a highly discerning attitude toward AI adoption. ET14 stated, "I refuse to adopt every new AI gimmick; I only use tools that have proven pedagogical merit."

Finally, participants across both groups highlighted disparities in access to AI tools and resources as a significant structural barrier. NT5 lamented, “Not all schools have the infrastructure to support AI integration. It creates inequities.” ET6 added, “Even within institutions, some teachers lack access to essential tools. This imbalance hinders progress.” These observations draw attention to structural inequalities, calling for equitable distribution of technological resources to ensure inclusive implementation.

4.5 The Imperative for Professional Development

A shared consensus emerged across the outcome space regarding the urgent need for professional development programs focused on AI. NT7 advocated, “We need workshops and courses to help us understand and implement AI effectively.” ET3 agreed, stating, “Without proper training, we’ll struggle to maximize AI’s potential or address its limitations.” Expanding on this necessity, NT15 emphasized, "We need continuous, context-specific training, not just one-off workshops on how to use the software," while ET15 agreed, "Professional development must address the ethical and identity-related anxieties we face, not just the technical skills." These calls for action highlight the gap between current competencies and future demands, emphasizing the necessity of systemic investment in teacher education.

5. Discussion

This study investigated the perceptions of novice and experienced Iranian EFL teachers regarding the role of AI in shaping their professional identities, revealing both distinct and shared perspectives. To explicitly address the research question, how novice and experienced Iranian EFL teachers perceive the role of AI in shaping their professional identity, the findings reveal that AI acts as a catalyst for identity renegotiation, but the trajectory of this renegotiation is heavily mediated by career stage and sociopolitical context. Novice teachers perceive AI as an identity-affirming scaffold that accelerates their transition into the profession, albeit accompanied by existential anxieties. Conversely, experienced teachers perceive AI as an identity-disrupting force that threatens their hard-earned autonomy and epistemological authority, leading to selective resistance.

The subtle differences between these two groups highlight the complex interplay between career stage, technological familiarity, and sociocultural context in determining how AI is internalized and integrated into pedagogical practices (Li, 2025; Yazan, 2018). As a dynamic and multifaceted construct, teacher professional identity is continuously (re)negotiated through interactions with institutional structures, sociopolitical ideologies, and emerging technologies such as AI (Beauchamp & Thomas, 2009; Beijaard et al., 2004; De Costa & Norton, 2017; Lan, 2024; Liu & Fan, 2025; Zhao et al., 2022). These divergent perceptions can be understood through the lens of Language Teacher Identity (LTI) theory, which posits that identity is a dynamic, socially constructed negotiation between personal agency and contextual constraints (Beauchamp & Thomas, 2009; Yazan, 2018). For novices, AI serves as a "community of practice" surrogate, mitigating the peripheral status often experienced in early career stages (Pennington & Richards, 2016). For experienced teachers, the introduction of AI disrupts their established "sub-identities" (Kayi-Aydar, 2019), triggering an ecological tension at the micro-level of classroom autonomy and the macro-level of ideological resistance to foreign technological hegemony (Satvati et al., 2023).

Novice teachers largely viewed AI as a supportive scaffold that enhances their confidence and competence in the early stages of their careers. This aligns with previous research indicating that novice educators often rely on external tools to compensate for limited experience and to develop a sense of self-efficacy (Fairley, 2020; Huang et al., 2021; Lan, 2024; Pennington & Richards, 2016; Tschannen-Moran & Hoy, 2007). For example, participants reported that AI tools provided instant feedback, lesson planning support, and adaptive learning resources—functions traditionally fulfilled by mentorship or formal training programs. These findings resonate with Huang et al. (2021), who found that AI-assisted teaching environments can reduce anxiety among new teachers and facilitate more responsive instructional decision-making.

However, alongside this enthusiasm, novice teachers also expressed existential concerns about obsolescence and over-reliance. This duality reflects broader anxieties within the educational technology discourse, where AI is simultaneously seen as a liberating force and a potential threat to human-centered professions (Shneiderman, 2022; Zhao et al., 2022; Zhou & Hou, 2024). The fear that AI might replace teachers altogether indicates a need for critical reflection on how to balance technological integration with the irreplaceable elements of human interaction and empathy in language teaching (Chounta et al., 2022; Lan, 2024).

Interestingly, some novice teachers adopted a forward-looking perspective, perceiving AI not just as a tool but as a catalyst for professional growth and innovation. They framed themselves as pioneers willing to experiment with AI-driven pedagogies, suggesting a generational shift in attitudes toward technology-enhanced education. This aligns with Tsou et al. (2024), who observed that younger educators tend to embrace AI more readily, seeing it as an extension of their teaching rather than a disruption. Moreover, the use of AI as a motivational tool for student engagement highlights its potential to enhance learner-centered approaches, particularly when used creatively by digitally literate educators (Chounta et al., 2022; Kerimbayev et al., 2023).

In contrast, experienced teachers exhibited higher levels of skepticism and resistance toward AI, often questioning its relevance and effectiveness in established pedagogical frameworks. Their reluctance to adopt AI-based tools reflects a broader trend observed in studies of veteran educators, who often prioritize experiential knowledge and interpersonal connection over technological efficiency (Beijaard et al., 2023; Fairley, 2020; Pennington & Richards, 2016; Tschannen-Moran & Hoy, 2007). Many participants voiced concerns that AI undermines the core values of teaching, such as intuition, empathy, and cultural responsiveness, that cannot be replicated by algorithms (Lan, 2024; Yazan, 2018).

Furthermore, experienced teachers frequently highlighted the threat AI poses to their professional autonomy and expertise. Statements about AI taking over grading, lesson planning, and even content delivery suggest a deep-seated fear of deprofessionalization, a concern echoed in global discussions on the impact of automation on education (Zhang & Hwang, 2023; Zhao et al., 2022). This sentiment is particularly pronounced in contexts like Iran, where foreign language education is deeply politicized and localized, making any form of external interference, including algorithmic systems, viewed with suspicion (Davari & Aghagolzadeh, 2015; Rashidi & Hosseini, 2019; Rassouli & Osam, 2019).

Despite this resistance, some experienced teachers demonstrated a pragmatic approach, advocating for a balanced integration of AI that complements rather than replaces traditional methods. This selective adoption mirrors the ecological perspective proposed by Satvati et al. (2023), which emphasizes the importance of contextual adaptation when introducing digital innovations into education. The emphasis on critical evaluation before adopting AI tools further underscores the mature pedagogical stance of seasoned educators, who are less likely to be swayed by trends and more focused on maintaining educational integrity (Varghese et al., 2016).

Several themes emerged across both groups, reflecting shared concerns and opportunities. Chief among these was the recognition that AI functions as a double-edged sword, offering benefits such as efficiency, personalization, and innovation, while posing risks related to depersonalization, dependency, and inequity. This dual nature of AI has been documented in various educational contexts, reinforcing the idea that its impact is contingent upon how it is implemented and regulated (Chounta et al., 2022; Flores & Swennen, 2020; Lan, 2024; Robson, 2018; Zhao et al., 2022).

Another significant commonality was the perception that AI profoundly influences professional identity. Both novice and experienced teachers acknowledged the need to redefine their roles in light of technological advancements, transitioning from knowledge transmitters to facilitators, curators, or co-learners in AI-mediated environments (Derakhshan & Ghiasvand, 2024; Goto, 2021; Lan, 2024; Zhao et al., 2022; Zhou & Hou, 2024). However, the extent to which this redefinition was embraced varied considerably, with novice teachers more inclined toward reinvention and experienced teachers more cautious in their adaptation.

Additionally, both groups emphasized the urgent need for professional development initiatives focused on AI literacy. This finding supports earlier calls for systemic investment in teacher education to ensure equitable access to digital tools and pedagogical strategies (Eslamdoost et al., 2020; Lan, 2024; Mohammed & Nell Watson, 2019). Without adequate training, teachers may struggle to harness AI’s full potential or address its ethical and pedagogical challenges effectively.

Finally, concerns about the erosion of teacher-student relationships due to increased reliance on AI were prevalent across both groups. This echoes recent scholarship highlighting the irreplaceable value of human interaction in fostering trust, motivation, and cultural awareness, particularly in language education (Chounta et al., 2022; Derakhshan & Ghiasvand, 2024; Yazan & Lindahl, 2020). Thus, while AI can enhance certain aspects of instruction, it must not come at the expense of relational pedagogy, which remains central to effective teaching.

5.1 Implications for Theory Development

This study extends the theoretical understanding of LTI in the AI era by demonstrating that technological integration is not a neutral pedagogical shift but a profound identity negotiation process. Theoretically, it bridges the gap between ecological perspectives of teacher identity (Satvati et al., 2023) and technology acceptance models, illustrating that emotional and identity-based tensions are primary predictors of AI adoption, overshadowing mere technical proficiency. Furthermore, by contextualizing these findings within Iran’s ideological landscape, the study contributes to the decolonial discourse in applied linguistics, highlighting how non-Western EFL contexts experience AI not just as a tool, but as an ideological artifact that necessitates unique identity preservation strategies.

6. Conclusion

This study targeted AI in education by examining its impact on teacher professional identity from the perspectives of novice and experienced EFL teachers. The findings reveal that while both groups acknowledge AI’s transformative potential, they differ significantly in their emotional responses, readiness for adaptation, and conceptualizations of their evolving roles. Novice teachers generally view AI as a source of empowerment and innovation, whereas experienced teachers exhibit greater skepticism and resistance, rooted in concerns about autonomy, tradition, and the irreplaceable human element in teaching.

Practically, the findings dictate that teacher education programs must move beyond mere technical training. Policymakers and institution leaders must design differentiated, context-sensitive professional development that addresses the emotional and identity-related anxieties of both groups. For novices, training should focus on maintaining pedagogical agency and avoiding over-reliance, framing AI as a collaborative partner rather than a replacement. For experienced teachers, professional development should adopt an empathetic, peer-led model that validates their experiential knowledge, demonstrating how AI can augment rather than replace their human-centric pedagogical strengths. Moreover, efforts should be made to preserve the relational dimensions of teaching, ensuring that AI serves as a complement rather than a substitute for human interaction.

The study also highlights the importance of contextualizing AI integration within the specific sociocultural and political realities of each educational setting. In some EFL contexts like Iran, where English language teaching is ideologically charged and technologically constrained, the adoption of AI must be approached with sensitivity to historical, religious, and institutional dynamics (Davari & Aghagolzadeh, 2015; Rashidi & Hosseini, 2019). 

While this study provides foundational insights, several avenues for future research emerge. First, longitudinal studies are needed to track how these identity negotiations evolve as teachers gain more prolonged exposure to AI tools. Second, future research should investigate the students' perceptions of AI-mediated instruction and how student feedback loops impact teacher identity. Third, comparative studies across different centralized and decentralized educational contexts would help isolate the specific impact of ideological constraints on AI adoption. Finally, observational studies examining the actual classroom enactment of AI tools, rather than self-reported perceptions, would provide a more comprehensive understanding of the theory-practice gap in AI integration.

This study is not without limitations. The sample, while purposefully selected, is relatively small and limited to a specific demographic of highly educated EFL teachers in Tehran, which may affect the transferability of the findings to rural or less privileged educational contexts in Iran. Additionally, the reliance on self-reported data through interviews and reflections may be subject to social desirability bias, where participants might overstate their technological adaptability. Future studies should incorporate classroom observations and artifact analysis to triangulate self-reported identity claims with actual pedagogical practices.

Finally, AI presents both unprecedented opportunities and formidable challenges for EFL teachers. Its integration into the profession requires a balanced, informed, and inclusive approach—one that respects the complexities of teaching identity while embracing the possibilities of innovation. By fostering dialogue between educators, researchers, and policymakers, we can navigate this transition thoughtfully, ensuring that AI enhances rather than diminishes the art and science of language teaching.