How AI Disrupted Our Human Research: Lessons Learned from Africa

While AI-powered tools like TLDRthis and Inciteful offer convenient solutions for summarizing articles and finding relevant sources, the question remains: can humans be replaced in qualitative research?

Despite advancements in AI, human researchers possess unique abilities such as critical thinking, empathy, and contextual understanding that are crucial in qualitative research. Human researchers can interpret nuances, navigate complex social dynamics, and generate insights that AI may struggle to capture. Additionally, qualitative research often involves subjective experiences and emotions that require human interpretation and understanding. While AI can augment qualitative research by automating certain tasks and providing additional insights, it is unlikely to fully replace the role of human researchers in qualitative research endeavors.

It seems like your research into mobile dating during the COVID-19 pandemic led you to encounter the intersection of AI and human responses. Your aim was to understand how people navigate mobile dating amidst changing pandemic conditions in New Zealand. As part of your research, you prompted participants to share stories in response to hypothetical scenarios.Over 2021 and 2022, you received diverse and interesting responses from 110 participants recruited via Facebook. Each participant was compensated with a gift voucher for their time. These responses likely provided valuable insights into the social dynamics and individual experiences surrounding mobile dating during a time of significant societal change.

Participants shared vivid accounts of characters navigating the unique challenges of “Zoom dates,” grappling with disagreements over vaccination statuses and mask-wearing, and delving into passionate love stories filled with surprising details. Some participants even broke the fourth wall, addressing us directly to express frustration with word limits or prompt quality.These responses painted a rich tapestry of experiences, capturing the spectrum of emotions felt during online dating amidst lockdown boredom and loneliness, as well as the excitement and heartbreak of seeking love during the pandemic.

However, what stood out the most was the reminder of the unpredictable and colorful nature of human engagement in research. Participants took us on unexpected journeys and offered unsolicited feedback, showcasing the dynamic and sometimes irreverent nature of human participation.Yet, in our latest round of study conducted in late 2023, we noticed a clear shift in the 60 stories we received.

This time, many of the stories we received felt unnatural, with stilted word choices and overly formal language. Each narrative seemed to carry a moralistic tone, prescribing what one “should” do in certain situations. Upon employing AI detection tools like ZeroGPT, we concluded that participants, or even bots, were using AI to generate story responses, possibly in pursuit of the gift voucher with minimal effort.Contrary to claims suggesting that AI can adequately replicate human participation in research, we found AI-generated stories to be lacking in authenticity. This reinforced the importance of basing social research data on lived experiences.

The primary threat to human research may not be AI itself, but rather the underlying philosophy driving its use. Claims about AI’s ability to replace humans often originate from computer scientists or quantitative social scientists who measure human reasoning or behavior through quantifiable metrics. This approach risks oversimplifying human experiences to fit within computational frameworks.In contrast, as qualitative researchers interested in the nuanced, emotional, and lived experiences of individuals, we value the complexities of human interactions. We were drawn to the diverse range of emotions and challenges participants originally shared about online dating during the pandemic.

In our experience, AI poorly simulated these nuanced experiences. While some may argue that generative AI is inevitable or can offer valuable research tools, we believe that theoretically-driven, qualitative social research is best equipped to detect and mitigate AI interference.The presence of AI as an unwanted participant underscores the need for academic institutions to develop policies and practices to support researchers navigating the evolving AI landscape. Regardless of one’s theoretical orientation, addressing the involvement of AI is crucial for understanding human perspectives and experiences in social research.

Ultimately, the limitations of AI serve as a reminder of the irreplaceable value of human insight and empathy in research endeavors.Alexandra Gibson, Senior Lecturer in Health Psychology, Te Herenga Waka – Victoria University of Wellington

Alex Beattie, Research Fellow, School of Health, Te Herenga Waka – Victoria University of WellingtonThis article is republished from The Conversation Africa under a Creative Commons license. Read the original article.