1. Objective and scope
This guidance defines the principles and practices by which artificial intelligence can be used in the analysis of qualitative data in theses. It applies to supervisors and students and is suitable for, among others, content analysis, thematic analysis, discourse and narrative approaches.
2. Principles
Transparency: use of AI must be reported transparently (tool, version, task, impact). Failure to report may be interpreted
Responsibility: the student must demonstrate their own analytical contribution; AI is a support, not a substitute.
Data protection: personal data and sensitive content must be anonymised before use with AI or processed in approved environments.
Academic freedom: the supervisor has the right to allow or restrict use with justification.
3. Process and roles
The supervisor, together with the student, defines the goals, boundaries and timing of AI use.
The student produces their own analysis before leveraging AI and documents all steps.
The student describes AI use in the thesis using the form ‘plan for AI use in the thesis’ (link) and states the planned use in the thesis/research plan.
4. Permitted and prohibited uses
Permitted: brainstorming alternative classifications, consolidating codes, reflecting on interpretation, structuring reporting (not inventing content).
Prohibited: using AI as the primary analysis method, inputting identifiable data into unsecured services, presenting AI outputs as one’s own work.
5. Data protection and ethics
To process material with AI, participants must be informed using the form ‘Notice to thesis participants’ (link).
To process material containing personal data with AI, participants must be informed in a data protection notice (link)
Anonymise the material before use or use systems approved by the organisation.
Do not input materials related to special categories of personal data or confidential information into public services.
Keep a log: what data was input, when and for what purpose.
6. Evaluation and quality assurance
The student must present a critical comparison between their own analysis and the AI’s suggestions (confirming/deviating findings).
Justify choices: why was a particular AI suggestion accepted or rejected?
Include the study’s delimitations and limitations, including the “black box” nature of models and potential biases.
7. Reporting requirements
Describe tools and versions, tasks, type of inputs (e.g. summarized excerpts), evaluation criteria and impact on conclusions.
Attach example prompts and a short log as an appendix, if possible without risking data protection.
8. When is use restricted?
When data protection or contractual terms do not allow the processing of the material.
When learning objectives require the student to carry out the analysis entirely without AI.
When the method (e.g. Grounded Theory) emphasizes the researcher’s subjective interpretation and continual comparison.
9. Usage examples (with sample prompts)
Example 1 – Reflecting on themes: “I have coded the following anonymous excerpts [X]. Suggest up to 3 alternative themes and justify by referring to the excerpts.”
Example 2 – Merging codes: “Here is a list of my codes [Y]. Suggest ways to merge overlapping codes without losing meaning.”
Example 3 – Triangulation: “Compare my analysis [summary] and the following AI-generated suggestions. What are the most significant differences and what might they indicate about theoretical assumptions?”
10. Responsibilities and monitoring
The supervisor is responsible for the guideline and ensuring the student understands the instructions. The student is responsible for documenting AI use and justifying interpretations. Laurea monitors the implementation of practices and updates the guidance as necessary.