This supplementary dataset provides the methodological materials supporting the qualitative study on human oversight, task-related AI ownership, and discretion in AI-assisted knowledge work. It includes (1) conceptual definitions of the constructs Task-related AI Ownership and Discretion in AI-assisted Work, (2) the complete semi-structured interview guide used to investigate employees' experiences with large language models (LLMs) in organizational settings, and (3) the systematic coding framework applied during qualitative data analysis.

The coding framework documents the operationalization of the main analytical categories, including responsibility, ownership attribution, discretion, work archetypes, and task risk, together with coding rules, definitions, decision criteria, and illustrative participant quotations. The interview protocol captures employees' perceptions of responsibility, autonomy, human oversight, quality assurance, AI integration into work practices, and organizational governance of LLM use.

The material is provided to enhance methodological transparency, facilitate replication, and support future qualitative research on human-AI collaboration, responsible AI use, and oversight practices in organizational contexts.