What’s the Use of Qualitative Interviews?

Introduction

You might think that qualitative research interviewers can’t be replaced.

You might think that we need human participants to conduct interviews for the purpose of research.

You might already have taken a strong position in the debate about the use of artificial intelligence in qualitative inquiry, and by extension interview research in particular.

This paper explores these ideas and invites readers to reflect on and debate these issues further. We now know that generative artificial intelligence (GenAI) or intelligent agents can function as a designer of research studies, generating research and interview questions, and assisting with revising questions. Intelligent agents can be created to conduct interviews on behalf of researchers. GenAI can also serve as an interviewee—that is via Large Language Models or, with LLMs integrated into robots. Voice-to-text applications have long assisted researchers with transcriptions and translations. GenAI can function as both an analyst and reporter. If GenAI can assume all of these tasks in qualitative research, what then is the use of qualitative interviews?

I begin by looking at some of the ways GenAI is being used in interview research. I then briefly review historical moments in the use of interviews and the debates about interview research before concluding with my answer to the question “What’s the use of qualitative interviews?” Throughout this paper, I refer to forms of interviewing other than research interviewing. These include recruitment interviews and interviews accessible in public media such as journalistic interviews. Readers might argue that these are not relevant to research interviewing. I use these as examples for several reasons. First, although conducted for purposes other than research, recruitment and journalistic interviews still rely on the question-answer sequence as the basic building block of interaction. Second, digitized interviews are easily available for repeated viewing and analysis, and are massively used by people everywhere to learn about the world. Third, because we all live in a media-saturated society, researchers are also influenced by interviews as a fundamental way of constructing identity and getting to know the world.

Gen AI and Interview Research

First, how is generative AI being used in interview research? Conveo, Agent Interviews, Genway and AILyze are just some of the commercial tools used to create intelligent agents capable of autonomous interaction to conduct interviews. Whereas Conveo, Agent Interview and Genway market their products for business purposes, AILyze, a recent sponsor of the American Educational Research Association conference in 2026, includes academic researchers as an audience for their products. In addition to conducting interviews, these tools are touted as providing speed, pattern recognition, “real-time quality assessment and bias detection” and “global reach” (Agent Interview, https://www.agentinterviews.com/research, accessed February 26, 2026). Agent Interview, for example, asserts their purpose is:

To democratize research by creating AI that can conduct authentic, insightful conversations with people around the world, enabling organizations to understand their audiences at unprecedented scale. (https://www.agentinterviews.com/about, accessed February 25, 2026)

Genway asserts that use of the AI interview can multiply coverage for a “fraction of the cost” and “100 times the speed” (Genway.ai, accessed April 1, 2026).

The speed of possibilities is reflected by Conveo’s advertising. Conveo asserts that a study can be launched in “minutes” and customers can receive “analyzed results within hours,” “making it ideal for teams who need to move fast without sacrificing quality. The AI handles transcription, tagging, theme detection, and reporting automatically” (https://conveo.ai/, accessed February 25, 2026). Conveo reported $1.8 M in revenue in 2025 (with a 16-person team) and has raised seed funding of $5.3 million (https://conveo.ai/insights/conveo-raises-5-3m-to-revolutionize-market-research-with-ai-powered-video-interviews), indicating that the production of these types of interview tools are profit-generating, if not profitable. These tools provide ways to occupy the roles of interviewer, analyst, and reporter. Let’s look further at the role that AI can assume as interviewer.

How We Got Here

I invited readers to take a step back and look at how we got to this point in the history of interviewing, where we have the technology for autonomous agents to contribute to, if not take over all aspects related to the work of research interviewing. Let’s begin with the interview: Where did it come from? How did it develop?

The word interview is derived from the French “entrevoir,” meaning “to see each other.” This term was used to refer to the dialogues facilitated by diplomats as they negotiated with one another across cultural and political boundaries. In the Renaissance period, interviews were the province of elites who traveled abroad and engaged in diplomatic meetings. Natsvlishvilil (2013) points to the Georgian archbishop, diplomat, and travel writer Timothy Gabashvili, who lived in the 1700s as one of the first people to record interviews of people he met during his travels through the Ottoman territories and the Holy Land. In his book “Travels” (Gabashvili, 2013), he records the questions he asked of the people he met, along with the detailed stories they shared with him. Gabshvili was engaging in narrative inquiry long before the term was coined.

The first traces of recorded interviews are found in the newspapers of the mid-1800s, when journalists reached out to others to ask questions of elite individuals. Turnbull (1936) identifies Horace Greeley’s 1859 interview with Brigham Young, the head of the Church of Jesus Christ of Latter Day Saints, as the first interview published in an American newspaper (p. 279). In Turnbull’s view, this interview provided an excellent example of the genre, being “packed with real information” (p. 275). By the end of the 19th century, researchers such as Charles Booth (1840−1916) and W. E. B. DuBois (1868−1963) were conducting social surveys in London and Philadelphia, respectively. This involved asking people questions and using this information in their studies to map urban areas and learn about social groups.

By the early 20th century, the field of sociology had taken to conducting interviews for large-scale studies. One such study was The Hawthorne Study conducted by Elton Mayo with his research assistant Fritz Roethlisberger from Harvard University in the late 1920s (https://www.library.hbs.edu/hc/hawthorne/07.html#seven). This study included more than 21,000 interviews with workers at the Hawthorne Works factory. The researchers learned that workers were more forthcoming when open-ended questions were posed, and over time, interviews grew from 30 to 90 to 120 min in duration.

In the 1930s, another massive interview project was initiated under the auspices of President Franklin D. Roosevelt’s Works Progress Administration project during the Great Depression. This was the Federal Writers’ Project, which involved generating oral histories of thousands of people across the United States. In the following decade, the National Opinion Research Center (NORC) was established at the University of Denver in 1941 and later moved to the University of Chicago. The commercial opinion polling organizations Gallup and Roper had been established earlier, while NORC was established to support social science research. By the mid-1900s, interviewing had been massively taken up by opinion pollsters and commercial researchers, sociologists studying social phenomena, folklorists recording people’s life histories, and anthropologists examining different cultural contexts.

As a result of all this interviewing, there is a good deal of methodological work on interviewing published in the 1950s. For example, to take one discipline, sociology, this included articles by sociologists Everett Hughes, Mark Benney, and David Riesman (Riesman, 1956/1993, 1957/1993, 1964/1993; Riesman & Benney, 1956). By 1956, Benney and Hughes (1956) wrote,

Sociology has become the science of the interview, and that in two senses. In the first sense the interview has become the favored digging tool of a large army of sociologists. . . . by and large, the sociologist of North America, and in a slightly less degree in other countries has become an interviewer. The interview is his tool; his works bear the marks of it. (p. 137)

Readers will observe that all of the interviewers referred to so far have been male. It should be noted that many of the fieldworkers who conducted interviews for the Federal Writers’ Project were women, and in the field of anthropology, there were prominent women interviewers such as Zora Neal Hurston (Hurston, 1927, 2018), Hortense Powdermaker (Powdermaker, 1966), and Margaret Mead (Mead, 1961/1928), all of whom were conducting interviews in the early decades of the 20th century. Whereas University of Chicago sociologists such as Everett Hughes and his students, who included Howard Becker, Erving Goffman, and Anselm Strauss, engaged in long-term fieldwork incorporating participant observation and interviewing, survey research became more prominent from the 1960s onward (for more on this history, see Fontana & Frey, 1998).

If one takes a closer look at all this “interview research,” it is clear that across the 20th-century “interviews” as a term referred to a wide range of interactional formats. For example, gerontologist Belle Boone Beard’s (1898–1984) archived research papers reveal that much information about participants of her study of centenarians (Beard et al., 1991) was recorded via written surveys, some of which she appeared to have developed. Her archived files also include standardized surveys from other sources. In her book on centenarians, what she labeled as “interviews” included both standardized surveys and oral history interview formats.

From the 1970s through the 1990s, ethnographic (Spradley, 1979), feminist (Oakley, 1981), and narrative scholars (Mishler, 1986) began to open up interviewing to take in flexible forms that provided an alternative format to the standardized surveys that many scholars had been trained to conduct. A steady stream of texts on qualitative interviewing became available to qualitative researchers all over the world. Scholars also began to critique how qualitative researchers were theorizing and using interviews. For example, Jim Scheurich (1995) wrote that the “conventional, positivist view of interviewing vastly underestimates the complexity, uniqueness, and indeterminateness of each one-to-one human interaction” (p. 241). Scheurich was critiquing the modernist assumptions upon which Elliot Mishler’s (1986) narrative response to standardized interviewing was built. Scheurich argued that Mishler’s approach to narrative interviewing that aimed to “tame” and resolve the “ambiguities of communication” inherent in interviews ultimately fails (Scheurich, 1995, p. 243). Nevertheless, Scheurich did not reject interviewing outright. Rather, he called for researchers to take seriously the “fundamentally indeterminate” nature of interviewing (p. 249). He wrote: “The indeterminate totality of the interview always exceeds and transgresses our attempts to capture and categorize” (p. 249), and encouraged scholars to play around and experiment with interviewing in ways that “highlight the indeterminancy of interview interactions” (p. 250).

Alongside qualitative researchers’ applications of in-depth approaches to interviewing in the latter part of the 20th century was the rise of celebrity culture. Interviews were everywhere—with journalists such as David Frost, Mike Douglas, Barbara Walters, and Oprah and oral historians such as Studs Terkel and later Henry Louis Gates Jr. achieving fame as interviewers. So popular had the interview become in both social science research and in everyday society, that sociologists Paul Atkinson and David Silverman (1997) dubbed contemporary culture “The Interview Society.” What they mean by this is that the “interviews of various kinds are relied on disproportionately” as a way to understand the construction of the self (p. 309).

Silverman has critiqued qualitative researchers’ commonsense use of interviews to explore “‘perceptions,’ ‘motives’ or ‘experiences’” (p. 149), and their failures to account for the recipient-design of descriptions and attend to the interactional nature of interview data in analyses and interpretations (Silverman, 2017). He has drawn on Jonathan Potter and Alexa Hepburn’s (2012) suggestions to put forward recommendations to improve the credibility of interview research. Other scholars, such as Lisa Mazzei and Alecia Jackson (2012) and Elizabeth St. Pierre (2021), have sought to deconstruct concepts such as the interview and experiment with ways to “inquire differently” (p. 163).

For decades, methodological literature has included discussions of the problems of using interview accounts as a source of evidence about the world. First, there is recognition of the fallibility of self-reports to support assertions, since people design their accounts for specific recipients and may even misremember or intentionally mislead their recipients. Second, scholars have long observed the socially situated and co-constructed nature of the generation of descriptions (Holstein & Gubrium, 1995), which challenges decontextualized reports of findings from which interviewers’ contributions have been stripped (Potter & Hepburn, 2005, 2012). Third, the use of interview data as evidence relies on interviewers and interviewees having achieved mutual understanding, in spite of differences. Interactional research on interviewing has found that interviewers do not always recognize in the moment that they have not fully understood their participants, and speakers’ use of a non-native language complicates what can be expressed (Ohta & Prior, 2019). Fourth, interview researchers must still account for the problems of representing the other in their reports. This has been discussed for decades within the debate that we know as the crisis of representation.

Still, over the last 25 years, there has been no decline in the popularity of the interview method. Instead, there has been a profusion of innovations in interviewing practices. Not only has there been a proliferation of theorizations of interviewing, but there are innovations in interview formats, including object, photo, and graphic elicitation, collage making, along with mobile methods and incorporation of vignettes (Roulston, 2023). Scholars have also drawn on new technologies to conduct interviews. Text-based and online video conferencing applications and email have supplemented telephone interviews as interview modalities. Whereas a century ago, the standard context for an interview would have been face-to-face with an interviewer using a set of questions, perhaps with note-taking and later audio-recording, in 2026 an interview can be asynchronous or synchronous, involve one or more parties as both interviewer or interviewee, and include various elicitation devices that encourage participants to engage in multi-modal forms of expression.

To reiterate the question with which I opened this paper, now that technically we no longer need either interviewers or interviewees to generate information in response to interview questions: What’s the use of qualitative interviews? If researchers see their participants as receptacles of information that need to be mined, to use the metaphor put forward by Steinar Kvale (1996) 30 years ago, then an intelligent agent can easily generate that information cost-effectively with the minimum amount of effort, and at scale.

I argue for the retention of the human as both interviewer and interviewee. I believe that if we want to understand human experience, then we need to communicate with humans as interviewers, and we need to consult humans rather than the amalgamation of information used in training Large Language Models. To make this case I will turn to students I’ve had the good fortune to work with—for it is the next generation of qualitative scholars who I believe can best answer this question.

What’s the Use of Qualitative Interviews? A View From Students

I’ve been very privileged for the last 25 years to teach students who want to learn about qualitative inquiry. Interviewing is but one of the methods they explore in coursework. When students come to interviewing they tend to think that gaining information from interviews is easy. All one has to do is ask a question, and the participant will provide an answer. This is exactly how GenAI works—you ask a question, and it will supply an instantaneous response.

In teaching novices to interview, I find that they are quite often surprised at how hard it is to ask well-formulated questions that facilitate conversations. They find it challenging to attend fully to others’ stories. They are surprised when interviewees take unexpected detours when answering questions. As interviewers, when we work with both strangers and friends, we all must commit to engaging in an interactional dance if we want to learn from another. Let’s look at some examples.

One Final Example

In conclusion, I share one personal example to illustrate what happens when GenAI is integrated into interviews. Over a year ago, I was talking to a colleague about interviewing LLMs. I had been experimenting with this out of personal curiosity. Shortly after our meeting ended, he sent me a transcript of an interview he had conducted with me. ³ You might be interested in what AI “I” had to say about this question:

As you can see, this is not too far from what I would like to convey to sum up this paper. ⁴ However, let’s look at another of “my” responses in this interview.

This is not too far from the truth, and it is somewhat frightening to consider that this response was generated from information I have shared online. However, here we see AI veer into a hallucination:

Alternatively, I could see myself in a profession focused on social advocacy or counseling. The relational and human-centered aspects of my work—listening, understanding people’s experiences, and reflecting on how we construct meaning—are what I enjoy most. Working in a counseling or social work role, where the focus is on helping people navigate life’s challenges, might have aligned with those interests as well.

This certainly sounds reasonable. Truthfully, though, there has never been a moment when I considered social work or counseling as professions.

Conclusion

To sum up, I argue that qualitative researchers continue to pursue opportunities to achieve intersubjectivity with other humans—that is, how we, as humans, might come to know the world in common. Although GenAI can now assist you in multiple ways—through taking on the role of “interviewer,” through providing data as an “interviewee,” through formulating interview questions and providing feedback to people learning how to conduct interviews—there are many good reasons not to delegate these tasks to a machine:

Gen AI questions

Listens, transcribes and translates;

reports. Cheap reports.

Rather than outsourcing these tasks, I recommend that qualitative researchers continue to take a lead role. And if you decide to incorporate GenAI into the process of interview research, be mindful, ethical, and vigilant. Do not give your role as researcher, creative thinker, and interpreter away:

Humans feel, learn, grow,

connect, relate, disagree.

Gen AI consumes.

What happens as a result of interviews involving humans? Interviewees and interviewers can develop and practice the art of human connection. They can develop experience in understanding, caring, and practicing empathy with one another. We as humans can experience what it is to be truly heard by another in a personal encounter, rather than have our language treated as data for reduction to informational content that can be bought and sold for profit. We can be surprised. We can be challenged. We can disagree. We can also embark on personal, meaningful encounters with one another that can change the ways we think and act in the world.

And is that not the ultimate goal of our inquiries?