Can Generative AI Generate Open-Ended Research Interview Questions?

Can AI Generate Interview Questions That Encourage Participants to Tell Their Stories?

You are preparing a qualitative research interview and want questions that encourage participants to describe their experiences in their own words. Rather than drafting every question manually, you ask ChatGPT to generate open-ended questions from your research objectives.

The system produces questions beginning with "How," "What," and "Why." They appear suitable for qualitative research because they cannot easily be answered with yes or no.

But an open-ended grammatical structure does not guarantee a useful qualitative question. "Why do teachers resist generative AI?" invites an explanation, yet it assumes that teachers resist the technology. "What are the benefits and challenges of AI?" allows lengthy answers but may produce generic opinions rather than accounts of actual experiences.

How can researchers use AI to generate questions that genuinely support qualitative inquiry rather than merely look open-ended?

Yes, but Open-Ended Wording Must Serve the Research Inquiry

How to Generate Open-Ended Questions That Support Qualitative Inquiry

What Makes an Interview Question Open-Ended?

An open-ended interview question allows participants to formulate an answer using their own language rather than selecting from predetermined response categories.

For example, "Have you used generative AI in teaching?" primarily requests a yes-or-no answer. By contrast, "Can you describe your experiences with generative AI in teaching, if any?" allows participants to explain what happened and how they understood those experiences.

Open-ended questions may elicit descriptions, narratives, interpretations, or reflections. Their usefulness depends on the research purpose.

Not every closed question is inappropriate during a qualitative interview. Brief factual or screening questions may establish relevant background information. The concern is whether the main questions provide sufficient opportunities for participants to explain their experiences and perspectives.

Open-Ended Does Not Automatically Mean Nonleading

One of the most consequential mistakes in AI-generated interview questions is confusing grammatical openness with neutrality.

Consider the question "How has generative AI improved your teaching?" Participants can provide extended answers, but the question assumes that improvement has occurred.

A less presumptive alternative is "What changes, if any, have you noticed in your teaching since you began using generative AI?"

Even this question should be directed only to participants for whom the reference to AI use is appropriate. A study including nonusers would need to accommodate their experiences as well.

A question can be open-ended and leading at the same time. AI can help identify leading interview questions, but its suggestions require researcher evaluation.

Translate Research Questions Into Participant-Friendly Questions

Qualitative research questions often contain theoretical language that may be appropriate for a proposal but unsuitable for direct use during an interview.

For example, a study might ask: "How do university instructors negotiate professional autonomy in AI-mediated instructional practices?"

Asking participants to explain how they "negotiate professional autonomy" may introduce terminology that is unfamiliar or carries different meanings for different instructors.

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

Follow-up discussion could explore what influenced the decision and whether the instructor felt free to choose.

The Interview Protocol Refinement framework proposed by Castillo-Montoya (2016) emphasizes aligning interview questions with research questions while developing an inquiry-based conversation. This distinction helps researchers avoid simply copying proposal language into the interview guide.

AI can assist with this translation, but the researcher must determine whether the resulting questions will provide material relevant to the inquiry.

Different Qualitative Purposes Call for Different Questions

Open-ended questions are not interchangeable across qualitative approaches. A question that is useful for exploring a process may not be equally suitable for investigating the meaning of a lived experience.

These examples illustrate different questioning purposes rather than fixed question sets for particular methodologies. Researchers should select questions according to their epistemological orientation, methodological design, and the phenomenon under investigation.

Why Concrete Experience Questions May Be More Useful Than Generic Opinions

AI frequently suggests questions such as "What are the advantages and disadvantages of generative AI?" or "What challenges do teachers encounter?"

These questions may be relevant, particularly when the study concerns general perceptions. However, they can also encourage participants to repeat familiar public discussions rather than describe events they personally experienced.

For a study of instructional decision-making, asking "Can you describe the last time you considered using generative AI for a teaching task?" may provide a more specific account.

The researcher can then explore what happened, which alternatives were considered, and how the participant interpreted the outcome.

This does not mean that concrete questions are always superior. Some studies legitimately investigate general beliefs, professional values, or collective understandings. The question should match the phenomenon rather than follow a universal preference for anecdotes.

Avoid Questions That Combine Several Demands

An AI-generated question may be open-ended yet require participants to address multiple issues at once.

For example: "How has AI changed your lesson planning, assessment practices, relationships with students, and professional identity?"

Participants may focus on one component while overlooking the others, or struggle to organize a coherent answer.

A more manageable approach is to ask about a central experience and use optional follow-up questions to explore relevant aspects as they emerge.

For example, "Can you describe how generative AI has figured in your lesson preparation, if at all?" could be followed by questions about assessment or professional responsibilities when those topics are relevant.

The objective is not to eliminate complexity from the interview. It is to avoid placing unnecessary cognitive demands within a single question.

How Should You Prompt AI to Generate Better Open-Ended Questions?

A useful prompt provides the research question, participant group, qualitative approach, and relevant context. It should specify the kind of account the researcher hopes to invite without prescribing the participant's answer.

Requesting explanations helps researchers examine the intended contribution of each question. However, those explanations are not evidence that the questions will function as expected during actual interviews.

Do Not Turn Potential Findings Into Interview Questions

AI may generate questions around familiar themes such as efficiency, ethical concerns, resistance, institutional support, or professional identity.

These topics may be relevant, but their appearance in an AI-generated guide does not establish that they are important in the participants' experiences.

For example, asking "How did institutional resistance affect your AI adoption?" assumes that institutional resistance existed and influenced the participant.

A more open formulation might ask: "What role, if any, did institutional expectations play in your decisions about using AI?"

Researchers should distinguish topics worth exploring from findings that can only emerge through data collection and analysis.

This is especially important when avoiding generic AI-generated interview guides that reproduce familiar themes without sufficient grounding in the study.

Allow Participants to Disagree With the Question's Premise

Good qualitative questions should not force participants to accept experiences or conditions they have not encountered.

Suppose a study includes teachers who have chosen not to use generative AI. Asking every participant "How do you integrate AI into your teaching?" excludes their circumstances.

An opening question such as "What has your experience been with decisions about using generative AI in teaching?" may accommodate both users and nonusers, although further wording review may be needed for the particular population.

Follow-up questions can then be tailored to the participant's account.

AI can help adapt interview questions for different participant groups, but the researcher should preserve the study's intended lines of inquiry.

Use Follow-Up Probes Rather Than Overloading Main Questions

Open-ended questions do not need to anticipate every possible response. Their purpose is to initiate a relevant account that can be explored through attentive interviewing.

After asking a participant to describe a decision about AI use, the interviewer might ask what influenced the decision, what alternatives were considered, or how the experience was interpreted afterward.

These probes should respond to what the participant actually says rather than follow a mandatory sequence regardless of relevance.

AI can help prepare useful interview probes, but it cannot anticipate every meaningful direction a participant may introduce.

Review the Questions for Sensitivity and Ethical Suitability

An apparently neutral question may invite disclosure of sensitive experiences, professional disagreements, or identifiable information about other people.

For example, asking instructors to describe incidents involving unauthorized AI use could lead them to identify students or colleagues. The researcher should consider whether those details are necessary and how confidentiality will be protected.

AI can help flag possible concerns, but it should not be treated as an authority on participants' emotional reactions or institutional ethics requirements.

When relevant, researchers should evaluate potentially sensitive interview questions and provide participants with appropriate opportunities to decline or redirect discussion.

Why Pilot Interviewing Remains Necessary

A question may appear open, clear, and relevant during desk review but produce unexpected difficulties during an interview.

Participants may interpret terminology differently, provide only generalized answers, or find that a question combines too many topics. The interviewer may also discover that certain questions repeat material already discussed.

Kallio and colleagues (2016) identify pilot testing as part of systematic semi-structured interview guide development. Castillo-Montoya (2016) likewise includes piloting within the Interview Protocol Refinement framework.

AI can anticipate possible problems, but pilot testing an interview guide provides feedback from actual interview interaction.

Turning a Broad Research Question Into Open-Ended Interview Questions

The revised question does not prescribe a favorable or unfavorable interpretation of AI. It creates an opportunity for participants to describe their own decisions and the circumstances surrounding them.

The researcher still needs to listen carefully and pursue relevant experiences that were not anticipated when the question was drafted.

Common Mistakes When Generating Open-Ended Interview Questions With AI

Use AI to Open the Conversation, Not Predetermine Its Answers

When asking AI to generate open-ended interview questions, specify the qualitative research purpose and the type of participant account you hope to explore. Request alternative formulations and examine what each one assumes.

Prefer questions that are accessible to participants and allow meaningful variation in their answers. Avoid requiring participants to accept the researcher's interpretation before they can respond.

Review the complete set of questions for unnecessary repetition and gaps in the inquiry. Then pilot the guide and revise it based on actual interview experience.

The aim is not simply to produce questions that cannot be answered with yes or no. It is to create opportunities for participants to contribute meaningful accounts that the researcher can examine in relation to the study.

Before Using AI-Generated Open-Ended Interview Questions

Frequently Asked Questions About AI-Generated Open-Ended Interview Questions

AI Can Generate Open-Ended Questions, but Openness Requires More Than Grammar

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.