Liars and Cheats? Critically querying the ‘factual’ exclusion of imposter participation

In academia, researchers are debating the consequences of increasing Gen AI accessibility - Dustin Hafki and Kendra Briken discuss issues arising during a research project and how their own assumptions about expected responses gave them cause to think again.

The acceleration of Gen AI tool accessibility induced by ChatGPT has raised many new challenges, not least for institutions focussed on the creation of, and contribution to, knowledge in society.

AI tools impact on the very core of knowledge production, and concerns about epistemic harm caused by Gen AI based tools are increasingly raised. Critical feminist, race and disability scholars have rung alarm bells early on to evidence the systemic reproduction of power relations in AI. Equally, concerns around epistemic foundations – how do we know what we know? – are growing, while Gen AI tools seem to have the capacity to act as epistemic authority materialising ‘knowledge’.

Given the normalisation of digitally facilitated interview situations, what is at stake for knowledge that is based on understanding and sense-making? Analysing the interview situation, including the role of the researcher, and acknowledging the participant as a powerful agent with their own agenda, is a well-established part of interpretive qualitative research. In our recent research experience, we show how the use of critical reflexive interpretation can enrich data around suspicious participation, rather than rejecting ‘imposter’ interviews from the start.

In a recent research project, focussing on Care Workers and their lived experiences of having voice as employed by temporary work agencies, we decided to add a tiny incentive for participants - a mechanism well established for survey research and panel participants. During the interviews, we noticed multiple discrepancies around the legitimacy of claims that some participants made. A feeling of suspicion emerged in the process - something sounded ‘off’ but it was hard to pin down the why and how. It was obvious though that some participants joining the call as ‘care workers’ were unfamiliar with common terms and institutions in the sector. A care worker, for example, would recognise the acronym for the regulatory body ‘SSSC’ or the SVQ - qualifications legally required for starting work in care environments.

Other authors (e.g. Singh et al., 2025; Medero et al., 2025; Ridge et al., 2023) reported similar experiences, often in the field of healthcare. In their papers, the process of dealing with potential imposter participation is discussed as a straightforward process, and omits, as we argue, a critical reflection of the interview situation. In doing so, multiple ethical dilemmas appear: What are our assumptions about the interviewee, how can we - from a position of privilege - start suspecting and accusing our participants being fraudsters? What story do our accusations tell about ourselves, power relations, the impact on participants, and of course for the findings?

Many of our ‘suspicious’ participants were non-native migrant workers. It hence would be entirely plausible that they may misunderstand questions or would not be as familiar with specific terminology as expected. More importantly, they might have used Gen AI during the interview to verbalise their own thinking. Wouldn’t we agree that, just as with our students, one of the accepted use cases for Gen AI tools is their capacity as assistive technology?

We stepped back and started co-interviewing to gather a holistic look at the whole process. The post-interview chat revealed that throughout the interview, there was a continuous feeling that something was off; smaller issues that did not add up to our lived experience and expertise of doing interviews. Every micro-moment could be explained by a nervous participant, miscommunication, or simply by going against our expectations hence it could be an interesting finding. Falsely accusing and excluding those voices from the research would go against everything that we wanted to achieve. Simultaneously, including falsified or fabricated experiences might undermine the trustworthiness of the data.

How to act in line with researcher integrity, knowing that the responsibility of making a judgement in qualitative research by default is influenced by our own situatedness.

We felt ethically challenged to make accusations towards research subjects. After all, those participants have taken time out of their lives to participate in this project. What’s more, given the amount of rejections received, we were equally desperate and grateful for any worker willing to share their perspectives.

What about our own unconscious bias when incriminating ‘friendly and open’ participants – especially in relation to power differences between researchers and vulnerable migrant workers. How could we insert our own judgement on the authenticity of the experience of participants, especially where those life journeys and experiences are so different from our own experiences?

This highlights another key ethical dilemma, centred around the lack of conclusive evidence that can be brought forward towards participants perceived as fraudulent – qualitative interviews rely on interpretation guided by the idea of ‘understanding’ subjective perspectives. Hence, even if we could come to the ‘logical’ conclusion that we might have ‘evidence’ – the qualitative interview setting is not supposed to be assessed like a trial, and we are not to act as judges.

Instead of attributing ‘wrong terminology or meaningless sequences’, and seeking evidence, we started interpreting the interview situation following the ethics of interpretive qualitative research. Rather than framing the situation as an outlier and repeating the circle of excluding voices of potentially precarious working lives, our collective interpretation led to insightful explorative insights on how care work is framed even if imagined. We engaged with narratives that allowed for a speculative approach: What narratives emerge in these interviews? We started seeing communalities that allowed us to see a version of care work though potentially not based on the specific work experience we assumed, but nevertheless on lived experience stemming from lives lived with a need for care.

The concept of imposter participation so far relies on a positivist perspective that treats the interview like a fact-finding mission. While this approach might be legitimated in some settings, it contradicts the researcher integrity based on ethics that we align with critical theory and pragmatism. Rejecting the reflex to dismiss what doesn’t fit with our own frame, but instead following ‘outliers,’ expands our understanding and insights into otherwise rejected perspectives. Rejecting the suggestion to exclude participants based on ‘factual’ assessment in our case was a decision in line with our ethical and philosophical stances. There is much to disentangle around so called ‘imposter participation’ and Gen AI’s role in fabricating or supporting lived experiences as well as the interrelation of these phenomena in our case. For us, though, it opened glimpses into imaginatives of care we would have (dis)missed otherwise.