The rapid integration of Generative Artificial Intelligence (GenAI) in higher education has triggered a profound socio-professional disruption, particularly within Humanities and Social Sciences (HSS; Ferret et al., 2025). While institutional discourses often frame this transition through a techno-pedagogical lens, emphasizing innovation, adaptation, and assessment transformation, our longitudinal study reveals a deeper psychosocial crisis. We posit that GenAI does not merely challenge pedagogical methods but fundamentally destabilizes the professional identity of HSS scholars, whose expertise is grounded in the authenticity of writing as proof of critical thought.

This paper proposes a social psychology of “augmented” education, exploring how academic communities regulate the epistemic uncertainties introduced by GenAI. We define epistemic uncertainty as the radical incapacity to distinguish human from synthetic authorship, leading to a breakdown in both the didactic contract and interpersonal trust within the academic community. Unlike previous digital transformations, GenAI introduces a creative agent capable of producing scholarly outputs that challenge the traditional distinctiveness of academic expertise (Pasonen, 2025; Brynjolfsson et al., 2025).

Research Questions:

1. How does the intrusion of GenAI into student production affect the professional identity and epistemic beliefs of HSS scholars at micro (individual), meso (organizational), and macro (institutional) levels?
2. What social regulation strategies (resistance, securitization, re-territorialization, identity work) do academic communities develop to cope with this “augmented” uncertainty?
3. How do institutional mediation devices (training, guidelines) interact with grassroots regulation processes in facilitating or hindering collective identity renegotiation?

Theoretical Framework:


Our integrated framework articulates three complementary psychosocial dimensions:

1. Identity Process Theory (Breakwell, 1986; Vignoles et al., 2011)

We analyze the crisis through threatened identity principles. Following Endacott (2021) and Pasonen (2025), we hypothesize that GenAI attacks two core dimensions: the distinctiveness of the scholar (if AI writes like a human, what constitutes scholarly expertise?) and self-efficacy (loss of evaluative control). The demand for detection software, despite its technical unreliability, is interpreted not as naive technological solutionism but as a coping strategy to restore professional integrity and boundary work (Nelson & Irwin, 2014).

2. Social Representations of Professional Knowledge (Moscovici, 1988; Jodelet, 2008)

We investigate the conflict between competing representations: the institutional view of AI as a “cognitive partner” requiring pedagogical adaptation versus practitioners’ representation of AI as a “viral threat” to the central nucleus of their profession being the evaluation of authentic authorship. This polarization reflects what Wallin et al. (2022) identify as the dual potential of digitalization to either support or undermine professional identity.

3. Collective Efficacy and Identity Work (Bandura, 2000; Watson, 2008)

We examine how the academic community shifts from individual distress (“paranoia,” isolation) to collective identity work and normative reconstruction (Pasonen, 2025). By creating new local norms (oral exams, hybrid writing assignments, ethical charters) and deliberative spaces (general assemblies), the community attempts to rebuild “vicarious mastery” over technological disruption. This transition mirrors the movement from professional identity flux to strategic re-territorialization documented in creative knowledge work (Golgeci et al., 2025).

Objectives:


This study moves beyond the binary “pro-innovation versus anti-cheating” debate to document the hidden psychosocial costs of digital transition. It highlights the specific vulnerability of HSS disciplines, where writing style is inextricably linked to critical thinking, and proposes a model of “augmented professionalization” that integrates epistemic anxiety as a legitimate component of professional development rather than mere resistance to change. By adopting a multicausal perspective (ie. integrating micro-level identity threats, meso-level organizational mediation, and macro-level institutional discourses) we aim to provide actionable insights for supporting academic communities through this socio-professional transformation.

Research Context and Corpus This research draws on a unique longitudinal corpus (2024–2026) collected at a mid-sized European university. The study focuses on a specific faculty (Institute) encompassing Culture and Communication, Law, Literature, and Public Administration, disciplines where creative knowledge work and critical writing constitute the core of professional expertise (Donsbach, 2014; Pasonen, 2025). The dataset comprises: 1. Institutional Communications (N=23): Newsletters, training announcements, policy documents issued by the university’s pedagogical support unit, reflecting the “prescribed” discourse of AI integration. 2. Horizontal Professional Exchanges (N=37): Emails and discussion threads from the faculty mailing list, capturing spontaneous reactions, debates, and expressions of professional distress during critical periods (e.g., post-exam sessions, general assemblies). This dual corpus enables analysis of institutional trust ambivalence (Pasonen, 2025) with the tension between top-down techno-pedagogical framing and bottom-up identity protection strategies. Analytical Approach We employ a mixed-methods design combining lexicometric analysis and critical discourse analysis (CDA): Lexicometric Analysis (IRaMuTeQ) The entire corpus is processed using IRaMuTeQ to perform Descending Hierarchical Classification (DHC), identifying stable lexical classes representing distinct discursive worlds (e.g., “security/detection” versus “pedagogy/transformation”). Correspondence Factor Analysis (CFA) visualizes temporal evolution (2024 vs. 2026) and the opposition between institutional and peer-to-peer registers. This approach has proven effective in documenting professional identity flux in digital transformations (Wallin et al., 2022). Critical Discourse Analysis We qualitatively analyze “flashpoint events” (e.g., the February 2025 crisis following widespread suspicion of AI-generated assignments) to identify rhetorical strategies actors use to legitimize positions, appeals to “academic integrity,” “environmental impact,” or “institutional duty.” This analysis reveals the identity work (Watson, 2008; Pasonen, 2025) through which scholars negotiate professional boundaries under technological threat. Temporal Dimension The longitudinal design enables us to track the community’s evolution from individual distress (2024: demands for detection tools, expressions of powerlessness) to collective normative reconstruction (2026: co-created local policies, revaluation of oral assessments). This mirrors Pasonen’s (2025) finding that successful GenAI integration depends less on technical training than on facilitating spaces for collective identity renegotiation. Ethical Considerations All personal data have been anonymized. Analysis focuses on discursive patterns and professional roles rather than individual attribution. Preliminary analysis suggests three interconnected findings: 1. Identity Polarization Across Institutional Levels 2. The “Illusion of Control” Paradox 3. From Individual Distress to Collective Identity Work

IRaMuTeQ classification reveals a sharp divide between a “Techno-Pedagogical” cluster (institutional actors emphasizing innovation, training) and a “Forensic-Anxiety” cluster (HSS practitioners focusing on fraud detection, authenticity verification). This polarization confirms threatened professional identity seeking boundary-protecting mechanisms (Breakwell, 1986; Pasonen, 2025), reflecting what Pasonen terms identity of superior capability scholars’ need to assert human expertise. We observe persistent demand for detection tools even among scholars aware of their technical fallibility, suggesting these tools serve a primarily psychological function (restoring justice, efficacy...) rather than a practical one. This extends Pasonen’s (2025) observation of identity empowerment dynamics (ie. the dual-edged nature of technology that simultaneously threatens and potentially reinforces professional agency). The longitudinal perspective reveals a shift from individual complaints (2024–2025: “I can’t trust submissions anymore”) to collective normative reconstruction (2026: oral defenses, hybrid assessment charters). Institutional mediation plays an ambivalent role: technical training provides skills but often fails to address deeper epistemic anxiety, resolved only through peer deliberation in “safe” spaces (General Assemblies). These results support our “Regulation of Epistemic Uncertainties” model, suggesting successful AI integration depends on facilitating collective identity renegotiation rather than purely technical adaptation (Golgeci et al., 2025; Wallin et al., 2022). This aligns with Marzano et al.‘s (2003) demonstration that interventions targeting educators’ “mental set”, including professional beliefs and identity anchor, outperform technical strategies. Our multicausal approach prioritizes the psychosocial dimension as the foundation upon which technical competencies can be sustainably built. This study presents limitations. Our corpus limits generalizability across diverse academic cultures, excludes student voices, limiting understanding of the didactic contract’s bilateral dynamics, and timeframe captures early adaptation; longer-term studies are needed to assess whether collective normative reconstructionproves durable as GenAI evolves. Finally, while our mixed-methods approach reveals discursive patterns, ethnographic observation of actual classroom practices would strengthen ecological validity.

Bandura, A. (2000). Exercise of human agency through collective efficacy. Current Directions in Psychological Science, 9(3), 75-78. https://doi.org/10.1111/1467-8721.00064 Breakwell, G. M. (1986). Coping with threatened identities. Methuen. Brynjolfsson, E., Li, D., & Raymond, L. R. (2025). Generative AI at Work. The Quarterly Journal of Economics, qjae044. https://doi.org/10.1093/qje/qjae044 Budhwar, P., Chowdhury, S., Wood, G., Aguinis, H., Bamber, G. J., Beltran, J. R., Boselie, P., Cooke, F. L., Decker, S., DeNisi, A., Dey, P. K., Guest, D., Knoblich, A. J., Malik, A., Paauwe, J., Papagiannidis, S., Patel, C., Pereira, V., Ren, S., Rogelberg, S., Saunders, M. N. K., Tung, R. L. & Varma, A. (2023). Human resource management in the age of generative artificial intelligence: Perspectives and research directions on ChatGPT. Human Resource Management Journal, 33(3), 606–659. Endacott C. G. (2021).The work of identity construction in the age of intelligent machines[Doctoral dissertation, UC Santa Barbara]. UC Santa Barbara Electronic Theses and Dissertations.https://escholarship.org/uc/item/2kb6p061 Golgeci, I., Ritala, P., Arslan, A., McKenna, B. & Ali, I. (2025). Confronting and alleviating AI resistance in the workplace: An integrative review and a process framework. Human Resource Management Review, 35(2). Ferret, N., Abry, P., Cazabet, R., Magué, J.-P., Gournay, L., Thiburce, & Gabriel, P. (2025). Ce que l’IA fait à l’écriture scientifique : ethnographie participative d’un hackathon interdisciplinaire. Discourse and Writing/Rédactologie (accepted). Jodelet, D. (2008). Le mouvement de retour vers le sujet et l'approche des représentations sociales. Connexions, 89(1), 25-46. https://doi.org/10.3917/cnx.089.0025. Marzano, Marzano, & Pickering (2003). Classroom Management That Works: Research-Based Strategies for Every Teacher) Moscovici, S. (1988). Notes towards a description of social representations. European Journal of Social Psychology, 18(3), 211-250. Nelson, A. J., & Irwin, J. (2014). ”Defining What We Do-All Over Again”: Occupational Identity, Technological Change, and the Librarian/Internet-Search Relationship. Academy of Management Journal, 57(3), 892-928. Pasonen, J. (2025). Leadership and professional identity challenges in GenAI transformation. Master’s Thesis, LUT University. https://lutpub.lut.fi/bitstream/handle/10024/169093/Pro_gradu_Pasonen_Jarno.pdf?sequence=1&isAllowed=y Vignoles, V.L. (2011). Identity Motives. In: Schwartz, S., Luyckx, K., Vignoles, V. (eds) Handbook of Identity Theory and Research. Springer, New York, NY. https://doi.org/10.1007/978-1-4419-7988-9_18 Wallin, A., Nokelainen, P. & Kira, M. (2022). From Thriving Developers to Stagnant Self‑Doubters: An Identity‑Centered Approach to Exploring the Relationship Between Digitalization and Professional Development. Vocations and Learning, 15, 285–316.