Description
Disclosing exactly how AI tools were used in analysis, including model, version, prompts, tasks, validation and human oversight, so readers can judge the credibility of findings.
Journals and conferences now require disclosure of AI use, but disclosures are often vague ('ChatGPT was used to assist analysis'). A rigorous report states the tools and versions, the tasks delegated (transcription, coding, summarisation), prompts or prompt templates, settings, data handling and privacy measures, validation procedures and results (e.g. inter-rater agreement between AI and humans), human checkpoints, and limitations. This extends established reporting standards for qualitative research such as COREQ and SRQR. Rigour comes from making AI's role replicable in principle and from not listing AI as an author, consistent with COPE guidance, while taking human responsibility for findings.
Attributes
- Discipline
- Research Design & Rigor
- Loss type
- Latent
- Goal
- Communicate findings
- Stage
- Post-launch
- Data
- Analytical / expert
- Effort
- Hours
Steps
- Record AI usage throughout the project: tools, model versions, dates, settings and prompts.
- Describe each delegated task and its place in the analytic process.
- Report validation: sample sizes, agreement statistics (e.g. kappa per code) and how disagreements were resolved.
- Describe data protection measures (local versus cloud processing, anonymisation, provider terms).
- State human oversight checkpoints and who was accountable for interpretation.
- Provide prompts and schemas in supplementary materials and discuss limitations and possible biases.
Sources
- Local Whisper (privacy)
- data anonymisation pipeline
- IRR reporting standards
- publication disclosure norms
- COPE (2023) Authorship and AI tools: COPE position statement
- Tong, A., Sainsbury, P. & Craig, J. (2007) COREQ. International Journal for Quality in Health Care
- O'Brien, B. C. et al. (2014) Standards for reporting qualitative research (SRQR). Academic Medicine
- The Inquiry Series, C14 — Ethical Constraints & Epistemological Limits of AI-Assisted Qualitative Research
Also known as
- AI use disclosure
- Transparent reporting of AI in qualitative research
Categories
- Research Operations
- Research Design & Rigor