Description
An epistemological stance that treats AI as a measurement and processing instrument to be calibrated and validated, while interpretation, judgement and accountability for findings remain with researchers.
In qualitative research the researcher is traditionally the primary instrument, bringing reflexivity and accountability to interpretation. Framing AI as an analyst blurs responsibility and invites unexamined authority for fluent outputs. Framing it as an instrument means it is specified (model, version, prompts), calibrated (validated against human judgement on samples), scoped (used for tasks like transcription, segmentation and deductive tagging), and reported like any instrument, with its error characteristics. Interpretive moves such as abstraction, theorising, judging significance and recommending action stay with people who can be accountable. The stance is a governance principle that shapes pipeline design, checkpoints and reporting.
Attributes
- Discipline
- Research Design & Rigor
- Loss type
- Latent
- Stage
- Discovery, Definition
- Data
- Analytical / expert
- Effort
- Hours
Steps
- Write a short statement of AI's role for the project, listing tasks delegated and tasks reserved for researchers.
- Specify the AI instrument fully: tool, model version, prompts, parameters.
- Calibrate against human judgement on samples and report its error characteristics.
- Reserve interpretive decisions (themes, theory, significance, recommendations) for researchers and record who made them.
- Maintain reflexive memos on how AI outputs may have shaped researchers' thinking.
- Report AI's role as an instrument in methods sections and stakeholder briefings.
Sources
- Local Whisper (privacy)
- data anonymisation pipeline
- IRR reporting standards
- publication disclosure norms
- Lincoln, Y. S. & Guba, E. G. (1985) Naturalistic Inquiry
- Braun, V. & Clarke, V. (2022) Thematic Analysis: A Practical Guide
- The Inquiry Series, C14 — Ethical Constraints & Epistemological Limits of AI-Assisted Qualitative Research
Also known as
- Instrumental stance on AI in research
- Human interpretive authority
Categories
- Research Operations
- Research Design & Rigor