Analytical Framework: Voice, Reflexivity, and Dialogic Becoming

Analytical Framework: Voice, Reflexivity, and Dialogic Becoming

for: Dialoguing with the Algorithm: An Autoethnographic Study of Midlife Voice, Uncertainty, and Teacher Identity in a ChatGPT Exchange

by: Maria Lisak EdD (How to cite)

Rationale: As an aging, foreign educator working with generative AI, I use a framework that brings together identity, voice, and labor. These lenses help me ask what it means to teach—and to age—in a field shaped by algorithms and speed. I draw on aging and teacher identity, dialogic theories of voice, and critical perspectives on digital labor to make sense of the affective, epistemic, and ethical tensions in my exchange with ChatGPT. Here I share my literature review on voice, reflexivity, and dialogic becoming.

Voice, Reflexivity, and Dialogic Becoming

My inquiry is grounded in a view of teacher identity as relational, reflexive, and socially saturated, extending the overarching conceptual–theoretical framework on aging, identity, and labor in TESOL/EFL. This section develops the analytical lens I bring to five critical incident analyses, drawing on interconnected theoretical resources to examine how identity is dialogically constituted in AI-mediated encounters. I begin with Bakhtinian dialogism and voice, which foregrounds the co-presence of multiple perspectives in any utterance, and performativity, which frames identity as enacted through repeated discursive acts. I integrate perspectives on intra-activity and material–discursive practices to account for the entanglement of human and nonhuman agencies, and on collaborative authorship and voice ownership to explore shared meaning-making in teacher–AI exchanges. Insights from critical AI studies situate these interactions within broader sociotechnical structures, while attention to marginality and resistance illuminates how teachers negotiate power and voice from the periphery. Finally, principles of dialogic pedagogy and emergent selfhood frame identity as continually shaped through open-ended, ethically responsive engagement. Together, these lenses provide the theoretical anchors for my analysis, enabling a nuanced reading of how professional identity emerges through recursive positioning, response, and becoming in AI-mediated pedagogical spaces.

Bakhtinian Dialogism and Voice

Drawing from Bakhtin's (2010) concept of voice, I approach my prompts to ChatGPT not as isolated expressions but as utterances shaped by prior discourse, institutional norms, and ideological struggle. For Bakhtin, voice is never singular but always heteroglossic—a weaving together of multiple social languages, discourses, and perspectives that constitute any utterance. In the context of my AI exchange, my prompts carry traces of institutional discourse (university expectations, ESP pedagogy), professional identity performance (experienced educator, cultural mediator), and personal vulnerability (aging, foreign status, technological uncertainty).

Bakhtin's notion of addressivity is particularly crucial here: every utterance anticipates and responds to its addressee, shaping both content and tone. My interactions with ChatGPT demonstrate this addressivity in complex ways—I simultaneously address the AI as technological tool, potential collaborator, and implicit threat to my professional autonomy. The AI's responses, in turn, become part of the ongoing dialogic chain, requiring my interpretive response and positioning.

The concept of answerability (Bakhtin, 2010) further illuminates how identity work unfolds through AI dialogue. Each exchange becomes a moment where I must answer—ethically, pedagogically, culturally—for the positions I take and the authority I claim or cede. This answerability is not abstract but embodied and contextual, requiring me to account for my choices within the specific cultural and institutional context of Korean EFL education.

Performativity and Identity Constitution

Butler's (2009) theory of performativity provides essential scaffolding for understanding how teacher identity is constituted through repeated acts rather than expressed from some pre-existing core. Identity, for Butler, emerges through citational practices—the repetitive performance of norms that, through their very repetition, create the illusion of a natural, stable self. In my AI exchanges, identity performance becomes visible through shifts in discursive positioning: from tentative questioner to authoritative expert, from collaborative partner to resistant critic.

Butler's later work on precarity (2009) resonates particularly with my experience as an aging, foreign educator engaging AI tools. The precarious subject exists in a state of interdependence that is simultaneously enabling and threatening. My dialogue with ChatGPT embodies this paradox—the tool offers support and efficiency while simultaneously raising questions about my professional relevance and future viability. The performative work of identity thus occurs under conditions of uncertainty, where each interaction potentially reinforces or destabilizes my sense of professional self.

The iterative nature of performativity becomes particularly visible in AI-mediated encounters, where the rapid back-and-forth of prompts and responses creates multiple opportunities for identity work. Each prompt performs a slightly different version of "teacher-self," and each response from the AI requires new performative choices about how to position myself in relation to algorithmic authority.

Liminality and Threshold Experiences

Turner's (1969) concept of liminality illuminates the threshold quality of AI-mediated identity work. Liminal spaces are characterized by ambiguity, disorientation, and the suspension of normal social hierarchies and categories. My extended ChatGPT exchange functioned as such a liminal space—neither fully human nor machine-mediated, neither purely instrumental nor entirely relational.

Within liminal spaces, Turner identifies the emergence of communitas—a sense of shared humanity and equality that transcends normal social distinctions. While my AI dialogue cannot achieve communitas in Turner's full sense (given the AI's non-human status), it does create a space where normal power relations are suspended. The AI does not recognize my age, institutional status, or cultural positioning, creating a peculiar form of equality that is both liberating and unsettling.

The liminal persona that emerges in this space is characterized by what Turner calls "structural invisibility"—existing between and betwixt established categories. As an aging educator engaging AI, I occupied this liminal position: neither digital native nor complete technophobe, neither fully autonomous nor entirely dependent on algorithmic assistance.

Intra-activity and Material-Discursive Practices

Barad's (2007) concept of intra-activity provides crucial theoretical grounding for understanding how identity emerges through entanglement rather than interaction. Unlike interaction, which assumes pre-existing separate entities that come together, intra-activity suggests that entities emerge through their relational engagements. In my AI dialogue, "teacher" and "algorithm" do not exist as discrete entities that subsequently interact; rather, both emerge through their ongoing intra-active engagement.

Barad's notion of material-discursive practices helps explain how the apparently "immaterial" exchange with ChatGPT has material effects on identity formation. The digital interface, the speed of response, the visual formatting of text, the institutional context of syllabus design—all of these material conditions shape the discursive possibilities and constraints of the exchange. Identity work is thus never purely linguistic or cognitive but always materially embedded.

The concept of agential realism (Barad, 2007) challenges traditional humanist assumptions about agency by suggesting that agency is distributed across networks of relation rather than possessed by individual subjects. In my AI exchange, agency emerges through the intra-active entanglement of human intentions, algorithmic processes, institutional pressures, and cultural contexts. This distributed agency helps explain why the exchange felt neither fully under my control nor entirely determined by the AI.

Collaborative Authorship and Voice Ownership

The question of authorship in AI-mediated writing has been explored by scholars working at the intersection of composition studies and digital rhetoric. Brooke (2009) argues that digital writing practices require us to rethink traditional notions of individual authorship, moving toward understanding writing as assemblage—a heterogeneous network of human and non-human actors. My syllabus revision exemplifies this assemblage authorship, where the final product emerges through the entanglement of my pedagogical knowledge, student needs, institutional requirements, and AI-generated suggestions.

Johnson-Eilola and Selber's (2007) work on assemblage rhetoric provides additional theoretical grounding for understanding how authority and voice operate in collaborative human-AI writing. They argue that digital composing practices involve the strategic selection, arrangement, and modification of pre-existing materials rather than creation ex nihilo. This framework helps explain the revoicing work I engaged in throughout the AI exchange—not simply accepting or rejecting AI suggestions but actively recontextualizing them within my pedagogical framework.

The concept of distributed cognition (Hutchins, 1995), while originating in cognitive science, has been taken up by writing studies scholars to understand how thinking and composing occur across networks of tools, texts, and social actors. My AI exchange exemplifies distributed cognition in action: pedagogical knowledge emerges not from my individual expertise alone but through the dynamic interaction between my contextual understanding, the AI's pattern recognition capabilities, and the material constraints of the syllabus format.

Critical AI Studies and Voice

Recent scholarship in critical AI studies provides essential frameworks for understanding the power dynamics and ideological implications of human-AI dialogue. Noble's (2018) work on algorithmic oppression reveals how AI systems embed and amplify existing social biases, particularly around race and gender. In my exchange, the AI's tendency toward Western-centered pedagogical assumptions exemplifies these embedded biases, requiring constant vigilance and corrective intervention.

Benjamin's (2022) concept of the New Jim Crow of technology illuminates how supposedly neutral AI systems can reproduce and intensify existing inequalities. As an aging, foreign educator, my engagement with AI occurs within these broader patterns of technological stratification, where access to and comfort with digital tools becomes a marker of professional competence and institutional value.

Crawford's (2021) analysis of AI as planetary extraction provides another crucial lens for understanding the material and ethical implications of my AI engagement. The environmental costs of training and running large language models, the labor conditions of data workers who prepare training datasets, the geopolitical implications of AI development—all of these factors shape the ethical context within which my seemingly individual pedagogical choices occur.

Marginality and Resistance

Anzaldúa's (1987) concept of nepantla—the in-between space where identities are destabilized and reconfigured—provides crucial grounding for understanding how marginalized subjects navigate technological encounters. As a foreign educator in Korea, I occupy multiple nepantla spaces: between Korean and Western pedagogical traditions, between institutional expectations and student needs, between technological fluency and digital anxiety.

Cho's (2007) introduction of the Korean concept teum or tuem (틈)—meaning gap, crack, or breathing space—offers a culturally specific framework for understanding how marginality can become a source of creative possibility rather than simply exclusion. In my AI dialogue, moments of epistemic dissonance and cultural mismatch became teum—spaces where alternative approaches could emerge.

The work of decolonial AI scholars (Mohamed, Png, & Isaac, 2020) provides frameworks for understanding how AI systems can both reinforce and potentially disrupt colonial patterns of knowledge production. My resistance to AI-generated content that erased cultural specificity represents one form of decolonial praxis—the ongoing work of maintaining cultural integrity within technological systems designed according to Western epistemological assumptions.

Dialogic Pedagogy and Emergent Selfhood

Freire's (1970) concept of dialogic pedagogy takes on new dimensions in AI-mediated contexts. While Freire emphasized dialogue between humans as the foundation of critical consciousness, my experience suggests that human-AI dialogue can also become a site of critical reflection—not because the AI possesses consciousness, but because the encounter surfaces and makes visible my own pedagogical assumptions and ethical commitments.

The emergent self that develops through AI dialogue is characterized by what Shotter (2006) calls "withness" thinking—a form of responsive, relational understanding that emerges in the moment of encounter rather than being pre-planned or predetermined. This withness thinking became particularly visible in moments where AI responses surprised me, challenged my assumptions, or offered unexpected connections between ideas.

Synthesis: Identity as Ongoing Achievement

Together, these frameworks position identity not as a possession or achievement but as an ongoing relational accomplishment. In AI-mediated pedagogical encounters, this accomplishment occurs through the dynamic interplay of performative utterances, material conditions, cultural positioning, and algorithmic responsiveness. The teacher-self that emerges through ChatGPT dialogue is neither simply enhanced nor threatened by AI but reconstituted through the ongoing work of responsive, critical engagement.

This foundation supports my analysis of AI-mediated identity work as fundamentally dialogic—not in the sense of achieving mutual understanding between human and machine, but in the Bakhtinian sense of ongoing, unfinished, and ethically charged response-ability. The five analytical moments I identify in my ChatGPT exchange represent crystallizations of this ongoing dialogic work, revealing how professional identity emerges through the patient, recursive labor of positioning, responding, resisting, and becoming.

For this paper, click here.