From Writing to Research: Process-Based Authorship in AI-Assisted Literature Review, Data Analysis, and Scientific Inquiry

The first paper in this series established Process-Based Authorship as a normative-conceptual framework for evaluating intellectual contribution in AI-assisted writing, introducing five dimensions—steering, intellectual contribution, engagement, learning, and transparency—and proposing the Process Engagement Index (PEI) as a preliminary metric. That framework was developed primarily in the context of conceptual and theoretical writing. The present paper extends it to the domain of empirical research, where the stakes of intellectual delegation are arguably higher and the boundaries of legitimate AI assistance less clearly defined. It examines three specific empirical workflows—AI-assisted literature review, AI-assisted qualitative data analysis, and AI-assisted quantitative interpretation—and argues that in each case the Steering Principle is the critical safeguard of intellectual integrity. The paper then develops the Process Engagement Index in full, proposing a five-dimension instrument with preliminary operationalization criteria, verification mechanisms grounded in LMS metadata and interaction forensics, and a scoring rubric for educational and institutional use. It further introduces the concept of the Research Governance Function to describe the irreducibly human role in AI-assisted empirical inquiry: the capacity to formulate research problems, adjudicate between competing interpretations, assume methodological responsibility, and stand behind the validity of conclusions. The paper is presented as a normative-conceptual proposal and identifies psychometric validation of the PEI as the primary agenda for subsequent empirical research.