Can Generative AI Help Create a Research Interview Guide?

Can AI Help You Develop Questions for a Qualitative Research Interview?

You have formulated your qualitative research questions and decided to interview participants. Now you need an interview guide that will help you explore their experiences, perspectives, and interpretations without turning the conversation into a rigid questionnaire.

Generative AI tools such as ChatGPT can produce an interview guide within seconds. They can suggest introductory questions, organize discussion topics, formulate open-ended questions, and recommend follow-up prompts. The resulting document may appear complete enough to use immediately.

But a list of plausible questions is not necessarily a well-designed qualitative interview guide. Questions may reflect assumptions the researcher has not examined, overlook important features of the research context, or encourage participants to discuss what the AI system expects rather than what they have actually experienced.

The challenge is determining how AI can support interview preparation without restricting the exploratory and interpretive possibilities of qualitative research.

Yes, AI Can Help Develop an Interview Guide, but It Should Not Determine the Inquiry

How Generative AI Can Support Qualitative Interview Guide Development

What Is a Research Interview Guide?

A research interview guide is a document that helps an interviewer explore topics relevant to a qualitative study. It may contain opening questions, main questions, optional follow-up probes, transitions, and instructions concerning the conduct of the interview.

Its structure depends on the methodological approach and degree of interview standardization.

In a semi-structured interview, the guide generally identifies topics and questions that the researcher intends to address while allowing flexibility in wording, order, and follow-up discussion. The interviewer can pursue unexpected but relevant experiences rather than adhering mechanically to a predetermined sequence.

A structured interview typically places greater emphasis on standardized questions and administration. An unstructured interview may rely on broader topics and participant-led conversation rather than a detailed question list.

AI can help prepare materials for these approaches, but researchers should establish the interview design before asking the system to generate questions.

Begin With the Research Questions and Methodological Approach

The most important starting point is not the AI prompt. It is the relationship between the study's research questions, qualitative methodology, and the experiences the researcher intends to understand.

Consider a study examining university instructors' experiences of using generative AI in teaching. A researcher might be interested in how instructors incorporate AI into instructional preparation, negotiate institutional expectations, and interpret changes in their professional responsibilities.

Asking AI to generate questions about the benefits and challenges of AI could produce a familiar interview guide. However, this framing might overlook how instructors first encountered the technology, whether they chose to use it, or how their practices developed over time.

The researcher's methodological orientation also matters. A phenomenological study exploring lived experiences may call for different questioning from a grounded theory study investigating a social process or a case study examining practices within a particular institution.

AI can suggest questions, but it should not silently determine the theoretical assumptions or analytical priorities of the study.

What Can AI Contribute to an Interview Guide?

Generative AI can support several practical activities during interview preparation. Its suggestions are most useful when the researcher supplies the study's purpose, participants, context, and methodological orientation.

These contributions concern interview guide development. They do not establish that the questions will generate rich, relevant, or trustworthy qualitative data when used with actual participants.

Use Established Interview Development Frameworks

Researchers do not need to invent a development procedure simply because AI is involved. Established methodological frameworks can guide the process.

Kallio and colleagues (2016) conducted a systematic methodological review and proposed five phases for developing a semi-structured interview guide: identifying prerequisites, retrieving and using previous knowledge, formulating a preliminary guide, pilot testing, and presenting the complete guide.

Castillo-Montoya (2016) proposed the Interview Protocol Refinement framework, which emphasizes alignment with research questions, constructing an inquiry-based conversation, obtaining feedback, and piloting the protocol.

These frameworks offer complementary ways to examine interview guide quality. Neither requires AI, but both provide methodological foundations within which AI assistance can be incorporated.

A practical application would be to use AI during preliminary question formulation and refinement while retaining the established requirements for conceptual alignment, feedback, and pilot testing.

Why Interview Questions Should Not Simply Repeat Research Questions

A research question and an interview question serve different purposes.

Suppose the overarching research question is: "How do university instructors negotiate professional autonomy when incorporating generative AI into teaching?"

Asking participants, "How do you negotiate professional autonomy when incorporating generative AI?" may sound academically precise but could be difficult to answer. Participants may not use the term professional autonomy or interpret it in the way the researcher intends.

A more accessible interview question might be: "Can you tell me about a time when you had to decide whether or how to use generative AI in your teaching?"

The researcher can then ask how that decision was made, who influenced it, and whether institutional expectations played a role.

AI can help translate abstract research questions into accessible open-ended interview questions. However, the researcher should check whether the resulting questions preserve the inquiry without introducing assumptions about participants' experiences.

Ask About Experiences Before Imposing Interpretations

AI-generated interview guides sometimes begin with abstract evaluative questions such as "What are the advantages and disadvantages of generative AI in education?"

Such questions may be useful for some studies, but they can encourage generalized opinions rather than detailed accounts of experience.

If the study concerns how instructors actually use AI, a more productive opening might be: "Can you walk me through the last time you used generative AI while preparing a lesson?"

This question invites a concrete account. The interviewer can subsequently explore the participant's decisions, concerns, and interpretations.

The distinction matters because qualitative interviewing often seeks to understand how people experience and make sense of situations, not merely collect their opinions about a topic.

Researchers should nevertheless avoid assuming that experience-focused questions are universally superior. The appropriate question form depends on the study's methodological orientation and purpose.

Include Probes Without Turning the Guide Into an Interrogation

A main interview question may produce a brief answer that requires clarification or elaboration. Follow-up probes help the interviewer explore what participants mean and how they arrived at particular experiences or interpretations.

For example, after asking an instructor to describe an experience using AI for lesson preparation, the interviewer might ask:

• Can you tell me more about what happened?
• What influenced your decision at that point?
• Could you give me a specific example?
• How did you interpret that experience afterward?

These are illustrative prompts, not questions that must be asked in every interview.

AI can help develop possible interview probes, but effective probing depends on listening carefully and responding to the participant's account.

A guide containing dozens of mandatory follow-up questions may undermine the conversational flexibility that semi-structured interviewing is intended to provide.

Sequence Questions to Support a Natural Conversation

An interview guide should help participants enter the discussion without immediately requiring them to explain abstract or potentially sensitive experiences.

For many semi-structured interviews, it may be useful to begin with familiar background experiences, move toward the main topics, and conclude with reflection or an opportunity to add information.

However, the order should not be treated as universally fixed. Some methodological approaches deliberately begin with a broad narrative invitation, while others use more focused questions to establish the relevant context.

AI can suggest a conversational sequence, but researchers should evaluate whether the order supports the intended inquiry and allows participants to introduce unexpected topics.

Detailed decisions about sequencing interview questions should account for the research context and the anticipated interaction between interviewer and participant.

Watch for Leading Questions and Embedded Assumptions

An interview question can influence participants by assuming that an experience occurred or suggesting how it should be interpreted.

For example, "How has generative AI improved your teaching effectiveness?" presupposes improvement. Participants who experienced no change or negative consequences may feel encouraged to frame their answers positively.

A less presumptive question would be: "What changes, if any, have you noticed in your teaching since you began using generative AI?"

Even this revision requires care. The researcher should not assume that participants attribute every change to AI or that they necessarily use it.

AI can assist in identifying potentially leading interview questions, but its review cannot establish how participants will respond during an actual interview.

Do Not Let AI Make the Guide Too Generic

A common weakness of AI-generated interview guides is their reliance on broadly applicable questions. Prompts such as "What challenges have you experienced?" and "What recommendations would you make?" can be useful, but they may reveal little about the particular phenomenon unless connected to the research context.

Consider a study of instructors using generative AI under institutional restrictions. A generic guide might ask about benefits and challenges while overlooking how institutional rules are communicated, interpreted, negotiated, or inconsistently implemented.

These contextual details may be central to the study's contribution.

Researchers should provide AI with relevant background information and evaluate whether its suggestions address the actual phenomenon. The aim is to avoid an interview guide that is too generic for the research context.

At the same time, supplying context should not mean telling AI what findings to expect. The guide must remain open to accounts that contradict the researcher's initial assumptions.

How Should You Prompt AI to Create a Research Interview Guide?

A useful prompt identifies the study's purpose, research questions, qualitative approach, participants, and relevant context. It should also specify that the generated questions must remain open to unexpected experiences.

This prompt provides a clear development task without asking AI to determine what participants will say.

The researcher should still examine the guide critically, particularly where the generated questions reflect familiar narratives about technology adoption rather than the study's distinctive concerns.

What Should Be Included Beyond the Interview Questions?

A practical interview guide may contain more than a list of questions. Depending on the research design, it may also include introductory instructions, reminders about consent, transitions between topics, optional probes, and a closing invitation.

For example, the guide may remind the interviewer to explain the study's purpose, confirm the participant's willingness to continue, and obtain permission before recording.

However, an AI-generated consent script does not replace an approved consent process or institutional ethics requirements. Researchers must verify that the interview procedures are consistent with the approved study protocol.

The guide may also include reminders about confidentiality, voluntary participation, and the participant's ability to decline particular questions, as applicable to the study.

When participants may discuss identifiable colleagues, students, or institutional incidents, the researcher should consider how unnecessary disclosure can be minimized.

Can AI Adapt the Guide for Different Participant Groups?

AI can suggest variations in wording and examples when a study involves participants with different roles or experiences.

For instance, instructors, academic administrators, and instructional designers may encounter generative AI through different responsibilities. A question about classroom decision-making may need a different formulation for administrators who establish institutional policy.

Nevertheless, adaptation should preserve the study's intended lines of inquiry where comparison across groups is important.

Researchers should distinguish legitimate contextual adjustments from changes that introduce different assumptions or investigate entirely different phenomena.

These considerations become especially relevant when adapting interview guides for different participant groups.

Why Pilot Testing Still Matters

An interview guide can look coherent on paper but function poorly during conversation. Participants may misunderstand a question, provide only brief answers, or interpret a supposedly neutral prompt as suggesting an expected response.

Pilot interviewing can reveal whether the questions are understandable, the sequence feels natural, the planned duration is realistic, and the guide provides sufficient opportunities for elaboration.

Castillo-Montoya's Interview Protocol Refinement framework includes piloting as one of its four phases. Kallio and colleagues likewise identify pilot testing within their framework for developing semi-structured interview guides.

AI can simulate an interview to identify possible difficulties, but a simulated participant is not a substitute for observing how a real person responds.

Researchers should distinguish preliminary AI feedback from the evidence obtained through pilot testing an interview guide.

Preserve Researcher Reflexivity and Judgment

Interview guide development involves choices about which experiences deserve attention, how questions are framed, and what assumptions the researcher brings to the inquiry.

Generative AI may reproduce conventional categories, familiar theoretical language, or widely discussed concerns. These suggestions can be useful, but they may also narrow the range of experiences participants are invited to describe.

Researchers should ask whether an AI-generated question reflects the participants' likely language or the researcher's preferred interpretation. They should also consider which experiences the guide may unintentionally exclude.

Reflexivity involves examining how the researcher's position, expectations, and decisions shape the research process. AI assistance does not remove this responsibility.

Maintaining substantive researcher judgment in qualitative question design is particularly important when AI-generated suggestions appear methodologically convincing but are insufficiently grounded in the actual research context.

Using AI to Develop an Interview Guide About Teachers' Generative AI Experiences

The revised guide does not require participants to classify their experiences immediately as benefits or challenges. Instead, it invites accounts from which the researcher can explore how decisions were made and what those decisions meant to participants.

The AI system helped generate alternatives, but the researcher determined which questions were appropriate for the qualitative inquiry.

Common Mistakes When Using AI to Develop Interview Guides

Use AI to Prepare the Conversation Without Predetermining Its Direction

When creating an interview guide, first establish what the study seeks to understand and why interviewing is appropriate. Then use AI to propose questions, examine alternatives, and identify possible weaknesses.

Review each suggestion against the methodological approach and research context. Consider whether the question invites participants to describe their experiences or encourages them to confirm a predetermined interpretation.

Keep a record of important development decisions, particularly when AI suggestions are accepted, revised, or rejected. Such documentation can support methodological transparency and researcher reflexivity.

During actual interviews, remain attentive to participants' accounts. A guide should support the inquiry, not prevent the researcher from recognizing something important that was never anticipated during question development.

Before Using an AI-Assisted Research Interview Guide

Frequently Asked Questions About AI-Generated Interview Guides

AI Can Help Prepare the Interview, but It Cannot Anticipate Every Meaningful Answer

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.