[Paper] Role and Identity Work of Software Engineering Professionals in the Generative AI Era

Source: arXiv - 2602.18190v1

Overview

The paper Role and Identity Work of Software Engineering Professionals in the Generative AI Era examines how the rise of Generative AI (GenAI) reshapes not only the tools developers use, but also how they see themselves and perform “identity work” – the mental and social processes of forming, adapting, or discarding professional identities. The author argues that a software engineer’s role (e.g., developer, tester, DevOps engineer) is a crucial, yet under‑explored, factor that influences this identity work.

Key Contributions

• Role‑centric framing – Introduces the idea that different SE roles experience distinct identity‑work dynamics when GenAI is adopted.
• Literature synthesis – Reviews existing studies on GenAI adoption in SE and on identity work, highlighting gaps related to role differentiation.
• Research agenda – Proposes concrete research questions and methodological directions to investigate role‑specific identity work.
• Practice‑oriented outlook – Discusses how insights from future studies could guide the design of tools, training, and organizational policies that support smooth GenAI integration.

Methodology

The paper follows a position‑paper / literature‑review approach:

1. Systematic scoping of recent publications on GenAI in software engineering (e.g., code‑generation assistants, test‑case synthesis).
2. Thematic analysis of identity‑work literature from sociology and HCI, focusing on concepts such as role identity, professional self‑concept, and identity negotiation.
3. Cross‑mapping of the two bodies of work to surface where role‑specific identity issues are mentioned or omitted.
4. Derivation of a research agenda by identifying unanswered questions and suggesting suitable empirical methods (surveys, interviews, ethnography, longitudinal case studies).

The methodology is deliberately high‑level, aiming to set a research direction rather than report new empirical data.

Results & Findings

• Current research conflates roles – Most GenAI studies treat “software engineers” as a monolith, ignoring the nuanced ways developers, testers, and other specialists interact with AI tools.
• Identity work is already observable – Early reports show developers feeling “augmented” or “threatened” by code‑completion tools, while testers express concerns about AI‑generated test artifacts replacing manual expertise.
• Role influences perceived agency – Professionals whose daily tasks are highly creative (e.g., feature design) tend to view GenAI as a collaborator, whereas those in more procedural roles (e.g., regression testing) may see it as a potential replacement.
• Research gaps identified – No systematic investigation of how role‑specific workflows, career trajectories, or team structures mediate identity work in the GenAI context.

Practical Implications

• Tool designers should tailor UI/UX and affordances to the target role (e.g., richer explainability for testers, suggestion ranking for developers).
• Team leads & managers can anticipate identity‑related resistance or enthusiasm by mapping role responsibilities to the capabilities of the chosen GenAI solution.
• Training programs need role‑specific curricula: developers might focus on prompt engineering and model fine‑tuning, while testers could benefit from validation techniques for AI‑generated test cases.
• HR and career‑development pathways can incorporate “AI‑augmented” skill tracks, helping professionals re‑frame their identities positively rather than feeling displaced.
• Organizational policies around AI usage (e.g., code‑ownership, audit trails) should reflect the differing accountability expectations across roles.

Limitations & Future Work

• Conceptual scope – The paper does not present empirical data; its claims are based on literature synthesis, which may miss emerging, unpublished industry practices.
• Role granularity – While the author highlights developers and testers, many other roles (product managers, DevOps, UX designers) are only briefly mentioned.
• Future work – The author calls for mixed‑methods studies (surveys, longitudinal ethnographies, controlled experiments) that explicitly compare identity work across multiple SE roles, and for the creation of artefacts (toolkits, guidelines) that support role‑aware AI adoption.

Authors

• Jorge Melegati

Paper Information

• arXiv ID: 2602.18190v1
• Categories: cs.SE, cs.CY
• Published: February 20, 2026
• PDF: Download PDF