Can Generative AI Adapt an Interview Guide for Different Participant Groups?

Can AI Help You Interview Different Participant Groups Without Asking Everyone the Same Questions?

You are conducting a qualitative study involving university instructors, academic administrators, and instructional designers. All three groups have experiences relevant to generative AI in higher education, but their responsibilities and perspectives differ.

You have already developed an interview guide for instructors. Now you want to adapt it for administrators and instructional designers without creating entirely unrelated interviews.

Generative AI can quickly replace terms, change examples, and generate questions for each group. However, replacing "teacher" with "administrator" may not be sufficient. The groups may encounter the same phenomenon through different decisions, institutional responsibilities, and opportunities for action.

How can researchers use AI to develop participant-specific questions while preserving the substantive coherence of the study?

Yes, AI Can Adapt Interview Guides, but the Underlying Inquiry Must Remain Coherent

How AI Can Support Participant-Specific Interview Guide Adaptation

Why Different Participant Groups May Need Different Questions

Qualitative participants may occupy different positions within the phenomenon under investigation. Their roles influence what they can observe, which decisions they make, and how they interpret events.

For example, an instructor may describe deciding whether to use generative AI when preparing instructional materials. An academic administrator may discuss developing or implementing institutional policies. An instructional designer may explain how AI-related tools are incorporated into learning resources or faculty development activities.

These experiences may all contribute to understanding generative AI in higher education, but asking each participant to describe classroom decisions would not necessarily produce meaningful accounts.

AI can help identify these differences. The researcher must determine which distinctions are relevant to the research questions rather than assuming that every occupational role requires a completely different interview.

Distinguish the Shared Research Inquiry From Group-Specific Experiences

A multi-group qualitative study may investigate one overarching phenomenon through several participant perspectives.

Suppose the study asks: "How do different university stakeholders experience and negotiate institutional expectations concerning generative AI?"

The shared inquiry concerns institutional expectations and how people encounter or respond to them. However, the relevant experiences may differ across groups.

The questions are not identical, but they address related aspects of the same phenomenon. This can support an analysis that considers both shared patterns and differences associated with participants' roles.

Researchers should not assume that the resulting accounts are directly comparable in every respect. The analysis must recognize differences in participants' responsibilities and opportunities to experience the phenomenon.

Use a Core-and-Adapted Question Structure

One practical strategy is to organize the interview guide into common areas of inquiry and participant-specific questions.

A common question might ask: "Can you describe how generative AI has figured in your professional responsibilities, if at all?"

Subsequent questions can explore activities relevant to the participant's role. Instructors might discuss teaching decisions, administrators might describe policy responsibilities, and instructional designers might discuss supporting faculty or developing learning materials.

This approach can help preserve research alignment without imposing identical questions on participants whose experiences differ.

However, the core-and-adapted structure is a practical design option rather than a universal requirement for qualitative interviewing. Some methodological approaches may justify substantially different guides for different groups.

What Can AI Help Adapt?

AI can assist with several aspects of participant-specific question development, although the methodological significance of each modification differs.

AI-generated adaptations should be treated as candidates. A model may infer occupational responsibilities from general patterns that do not accurately describe the particular institution or participants.

Do Not Assume That Participants With the Same Role Have the Same Experience

Grouping participants by professional role can be useful, but it may also encourage overgeneralization.

For example, not every university instructor uses generative AI. Some may have extensive experience, others may have limited exposure, and some may have deliberately chosen not to use it.

Similarly, academic administrators may differ in their involvement with AI policy. An administrator responsible for academic planning may have different experiences from one whose responsibilities concern student services.

AI may generate questions based on stereotypical expectations about these roles. Researchers should avoid treating such expectations as verified descriptions of the participants.

Questions should allow participants to explain their actual responsibilities and experiences before more specific assumptions are introduced.

How Should You Prompt AI to Adapt an Interview Guide?

A useful prompt provides the original interview guide, research questions, participant groups, and relevant differences in their roles. It should also distinguish common topics from questions requiring substantive adaptation.

This prompt asks AI to preserve conceptual continuity rather than merely generate three independent question lists.

The researcher should verify the proposed questions against actual institutional arrangements and the methodological purpose of including each participant group.

When Does Adaptation Change the Research Question?

Some revisions go beyond making an interview question appropriate for another group. They introduce a different phenomenon.

For example, suppose a study investigates how institutional AI policies influence professional decision-making.

An instructor question about deciding whether to use AI for assessment may be relevant. An administrator question about how institutional policies are communicated may also contribute to the same inquiry.

However, replacing the administrator question with "What are the advantages of AI for university competitiveness?" shifts the focus toward institutional strategy and perceived benefits.

That topic may be interesting, but it is not automatically equivalent to the original inquiry.

Researchers should distinguish contextual adaptation from expansion or redirection of the research scope.

Does Every Participant Group Need Identical Questions for Comparison?

No. Qualitative comparison does not necessarily require identical interview wording.

Researchers may compare how different groups experience the same phenomenon while recognizing that their roles provide access to different aspects of it.

For example, instructors may describe how a policy affects their teaching decisions, while administrators explain how they interpret responsibility for implementing that policy.

These accounts may be analytically connected without being interchangeable.

However, if the study intends to compare responses to a particular question directly, maintaining sufficiently similar wording or a clearly shared topic may become more important.

The necessary consistency depends on the analytical purpose and qualitative approach. Researchers should not treat uniform wording as an automatic guarantee of meaningful comparability.

Can Different Guides Create Unequal Opportunities to Discuss Important Topics?

Yes. Adaptation can unintentionally make some topics available to one group but not another.

Suppose instructors are asked detailed questions about ethical concerns while administrators are asked only about implementation efficiency. The resulting differences in the interview data may partly reflect the questions asked rather than differences in participants' actual perspectives.

This does not mean every topic must be explored identically with every group. However, researchers should examine whether the guides provide appropriate opportunities to address the issues central to the research questions.

A cross-group question matrix can help identify topics that are omitted, emphasized unevenly, or framed differently without methodological justification.

AI can prepare such a matrix, but the researcher must determine whether the differences are defensible.

How Can AI Help Adapt Interview Probes?

Different participant groups may require different follow-up questions because their responsibilities and experiences differ.

After an instructor describes an AI-related teaching decision, the interviewer might ask what influenced the choice and how it affected instructional preparation.

After an administrator describes a policy decision, relevant probes might concern the circumstances of its development, the people involved, and how its implementation was understood.

AI can suggest interview probes for each group, but these should remain optional and responsive to participants' actual accounts.

The interviewer should avoid asking a prepared probe simply because it appears in the group-specific guide.

Review the Ethical Implications of Different Participant Roles

Participants' institutional positions may influence the sensitivity of particular questions and the consequences of disclosure.

For example, an instructor may hesitate to discuss disagreement with institutional policy, while an administrator may be concerned about revealing confidential deliberations or identifiable colleagues.

Researchers should consider whether questions invite unnecessary disclosure of restricted information, personal criticism, or details that could identify participants in small professional groups.

AI can help flag potential concerns, but it cannot determine the actual institutional risks or replace the study's ethical review.

Researchers should examine potentially sensitive interview questions in relation to each participant group rather than assuming that the same wording carries the same implications for everyone.

Why Pilot Interviews Should Include Relevant Participant Groups

An adapted question may appear appropriate on paper but prove confusing or irrelevant when asked to someone in the intended role.

For example, an AI-generated administrator question might refer to "AI governance committees" that do not exist in the institution. An instructional designer might interpret "AI integration" differently from the researcher.

Pilot interviewing with appropriately selected representatives can reveal these mismatches and identify whether the questions invite useful accounts.

Kallio and colleagues (2016) include pilot testing within their framework for developing semi-structured interview guides. Castillo-Montoya (2016) similarly emphasizes piloting as part of interview protocol refinement.

AI-assisted adaptation should therefore be followed by suitable review and pilot testing, particularly when substantial changes are introduced.

Document What Changed and Why

When a study uses multiple interview guides, researchers should maintain a record of common questions, participant-specific revisions, and the reasons for those differences.

A question matrix may include the research question, common topic, wording for each participant group, intended contribution, and relevant methodological notes.

This documentation can help readers understand how the different interviews contributed to the overall inquiry.

If AI materially assisted with adaptation, its role should be described accurately according to applicable institutional and publication requirements.

Maintaining researcher judgment in AI-assisted qualitative question design is particularly important when different guides could shape which perspectives become visible in the final analysis.

Adapting One Interview Guide for Instructors, Administrators, and Instructional Designers

The resulting guides contain different wording but remain connected to the same research problem.

The researcher can examine how the groups describe institutional expectations while recognizing that differences in their accounts may reflect both their roles and the questions used to explore those roles.

Common Mistakes When Adapting Interview Guides With AI

Adapt Questions to Participants Without Fragmenting the Research Inquiry

When interviewing different groups, use AI to identify where wording or topic emphasis may need adjustment. Begin with the shared research questions and then consider what each group can meaningfully contribute.

Do not assume that identical questions guarantee comparability or that different roles require completely separate inquiries.

Use a question matrix to document the shared topics, group-specific wording, and reasons for adaptation. Pilot the guides with relevant participants when feasible.

The objective is to allow different groups to contribute their perspectives on the same research problem without forcing them into questions that do not reflect their experiences.

Before Using AI-Adapted Interview Guides With Different Groups

Frequently Asked Questions About Adapting Interview Guides With AI

AI Can Adapt the Questions, but Researchers Must Preserve the Inquiry

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.