What Human Centric Qualitative Research Looks Like in the Hands of Artificial Intelligence

Published On: October 16, 2023By

Like most people, I’ve been following the fascinating yet frightening rise of AI. Whether it’s music, art, education, or testing its niche capabilities like resume editing and answering bizarrely specific questions. As I watched (and continue to watch) AI infiltrate different industries, I began to think about how it would affect my work in the marketing and qualitative research space. I optimistically like to think we are somewhat safe from the AI takeover being that the work we do is human, real people centric. With this notion in mind, I wanted to find a way to prove it.

After reading an article that Black Mirror creator Charlie Brooker had ChatGPT attempt to write an episode, I had the idea to have AI take a stab at what we do at KNow as a qualitative research agency. We cover various stages of research, everything from project design, recruitment, conducting fieldwork, analysis and more. Confident that a robot couldn’t possibly come close to producing work rooted in empathy and human connection, came KNow vs. AI – an experiment where we would go head to head with artificial intelligence to confirm or deny the importance of human touch in our work. 

KNow vs. AI Objective:

As a company that champions innovation, we knew that the rise of AI was not a trend to ignore or be afraid of, but an opportunity to see if/where it fits within our space and the work we do.

We started KNow vs. AI not as an experiment to see in what ways AI can replace what we do, but to support what we do. We’re testing various AI tools and capabilities to discover how they can assist our team in working more efficiently throughout the different stages of research. In this series we’ll be going head to head with various platforms and tools to uncover the gap or discrepancies between the work of humans and AI.

Through this series, we believe we’ll find that there is no replacing or replicating the power of the human touch, especially in our space of qualitative, human centric research. While conducting this experiment, we hope to adopt the AI tools that not only save our team time and money, but the tools that allow us to continue to provide our clients the best results.

To determine which research areas we wanted to test, we began thinking about pain points and time consuming areas throughout the research process. We decided for the first round of the experiment, we would have AI try some of the basics such as writing a project outline and hypothesizing the research outcome. Then taking it to the next level, we’d have AI write a discussion guide and screener based on the project objectives, as well as have it create an insight summary based on fieldwork transcripts. 

We strategically timed this experiment to happen in tandem with our annual Pro Bono Project. Our Pro Bono Project not only gifts research to an organization in need, it’s an opportunity for our team to try new tools, techniques and methodologies. As we compiled a list of AI platforms to try, we considered both research-based tools and open LLM tools. The issue with many open AI tools is they lack transparent data usage and storage policies, which conflicts with our commitment to participants’ privacy and client confidentiality. However, our Pro Bono Project recipient and participants granted us full consent to use their information from the research freely for this initiative. 

As our Pro Bono Project comes to a close, we’ve tested several AI tools throughout the process. Many of which are beta versions that our existing partners have been working on (whom we are working with as they continue developing), as well as new platforms chosen based on capabilities and relevancy to the scope of the project. Overall, it’s been an exciting process that we’ll continue to test as the AI space is constantly and rapidly evolving. With so many free resources available, you don’t want to fall behind. I highly encourage everyone to at least give it a try. Play around and test different platforms to figure out how AI can support what you do (not replace it)! Stay tuned, we’ll be diving into the AI results of the different research phases in upcoming blogs!