Researching AI video interviews from a human perspective

7 June 2022

A lot of work on Artificial Intelligence focuses on the benefits and costs for organisations, ignoring the human impacts of these technologies.

Our research, instead, focuses on the human experience of technology in organisations. We first focused on the impact that AI has at a human level when we were researching Automated Video Interviews (AVIs) and looking at their effects on young job-seekers. Mo (a pseudonym), a promising student from a less privileged background, told us that he had been interviewed by bots.

Mo felt lost. Did AVIs use face recognition? Did they interpret the words he used? Did they code his tone of voice into emotions?

Mo was even under the impression that the AVI bot had tracked his hand movements.

Our interest in this technology, therefore, started from a concern for the human. We wanted to understand how AVIs were perceived by young job-seekers, and whether AVIs might be contributing to an extra level of anxiety within the job seeking process. Consequently, our research questions were simple, borne from a  real issue, rather than the academic literature. We wanted to know: How are young job-seekers affected by AVIs? Most importantly, given Mo’s disorientation: How are students from less privileged backgrounds affected by AVIs?

Our human-driven focus, and the empathy that underpinned our research interest made drafting a grant proposal a very quick affair. We partnered with the Institute of Employment Studies (IES) and submitted our grant request to the HEIF Covid Recovery Fund at the University of Sussex. Having secured funding, our ongoing research started in May 2021. We recruited 20 participants from the University of Sussex, using opportunistic (email-based) sampling.

The human experience of AVIs

AVIs represent a fertile topic for research, as the technology is very new, and its use has increased much faster than the academic/practitioner literature on how to use it. Our first focus was therefore to understand how the technology worked. We familiarised ourselves with interview apps, websites, and literature. We even tested some face recognition software ourselves.

However, despite our increasing understanding of the technology, its complexity, and its limits, we still felt we were missing some important insights on the impact that the technology was having on humans. We decided to use human-centred methodology, by adopting a phenomenological approach to research. Our interviews zoomed into the subjectively lived experience of our young, job-seeking participants by asking about their perceptions. In speaking to research participants, we were almost agnostic to the technology, instead focusing our enquiries on what had happened to them before, during and after their AVIs.

We soon started to gather a clear view of the tensions that young job-seekers from low-privilege backgrounds face when subjected to AVIs. Four key themes started to develop:

1 AVIs Are Hard to Understand

First and foremost, we found that job seekers were confused about the type of interview they were being asked to undertake, and more specifically, the type of AVI involved. To help them with this, we were keen to build a visual representation of AVIs:

The Depersonalisation Spectrum: 1 Face-to-Face 2. Video Interviews 3. Asynchronous Video Interviews (AVIs) 4. AVI A.I.-Assisted 5. AVI A.I.-Led

Figure 1 – Our Depersonalisation Spectrum

2 Feelings of Humanity are Diminished

This lack of understanding from AVI job-seekers meant that, during the interviews, the candidates tried to perform in a rigid way (e.g. holding a fixed gaze, a fake smile, or unnatural posture; speaking with a monotonous voice; holding their hands still) to try to comply with a technology that they did not understand. Many told us that they felt they had to behave like robots.

3 AI Technology Is Glorified

This diminished humanity wasn’t always perceived as a negative thing, as some candidates believed that the technology was a more efficient way of screening than a human interaction, due to its perceived objectivity. Many considered the AI as superior to human cognition. Therefore, job-seekers saw their ‘robotic performance’ as an inevitable part of the recruitment experience and a necessary trade-off to enjoy a more objective application process.

4 AVIs Are Emotionally and Cognitively Exhausting

The constant unnatural performance that job-seekers reported, in front of a screen that did not return any cue of approval or disapproval, led to the overuse of emotional and cognitive resources. Many candidates told us they felt energy-depleted by this experience.

Helping students, employers and policymakers get to grips with AVIs

Our research has useful lessons for future jobseekers, employers and policymakers and we have produced a number of resources to help navigate this new technology.  We intend to continue with our phenomenological approach and to widen our research to other demographic groups, e.g., workers who are older, or more experienced. We are also keen to integrate our interviews with data gathering techniques that allow us to capture the human experience more directly, such as with self-reflective diaries and to study other kinds of HRM technologies, like electronic performance monitoring or algorithmic management apps. This kind of ‘human-centred’ phenomenological approach has the potential to provide an important perspective on the proliferation of digital technologies at work.

Media coverage

Our way of communicating the research was as important to us as the research itself. For example, we were very keen to use visuals, videos and illustrated reports to communicate the human-centred findings. This made our academic findings much more accessible and easy to understand. By carefully coordinating the research release with our press officer, the research got huge attention in October 2021: it informed an FT work column, resulted in a BBC News interview and a Harvard Business Review article.

Resources

For policymakers

Policy Brief

For employers

Employer Tool Kit designed with the IES, to be distributed through their networks to employers countrywide.

For students

Two videos explaining to students what AVIs are, and how they can be affected by this technology (see video 1 ‘Become Aware’ and video 2 ‘Stay Human!’)

Related event

Digit Debates: Automating job interviews – experiences and implications for job seekers. Register

Recent blogs

Share this: