By Michael Mermelstein, Executive VP/Partner, Nichols Research, Fremont, California, michael@nicholsresearch.com
Adapting to the Next Frontier—AI: Friend or Foe?
When we look back over the last three years, we realize how quallies and market researchers in general were not ready when the COVID-19 pandemic struck. Many of us were not aware of the online platforms that existed; we needed to learn the different ways and nuances of moderating online versus in person, and we needed to work out the kinks from group length, group size, tech checks, and so on to be successful. We adapted and have done well.
But now comes the next challenge to how we conduct qualitative research. Actually, it has been around for a number of years—but the technology, techniques, and therefore the uses continue to get better and morph into more helpful tools. I am talking about artificial intelligence, or AI. To date, much of what has been developed is based on machine learning, where computers are trained to recognize specific patterns and then use algorithms to continue to refine what is happening more quickly than humans can. AI is the simulation of human intelligence processes by computers; it can expertly process language and recognize speech patterns… and talk back to us, answer our questions, and even develop questions to ask.
Market research companies such as Quester and Remesh have been using machine learning in various ways for a number of years and continue to refine their capabilities. Now there is ChatGPT, which is built on an open AI platform. ChatGPT is based on natural language techniques with a bot that can respond to queries in a cohesive and human-like way.
Is AI a friend or foe of market research? It depends on how we adapt. Researchers say that AI will bring greater data accuracy, will increase the size of the market research industry, and will create or maintain more jobs than it destroys. This remains to be seen. But ultimately, AI will change the roles we need to perform (as opposed to our computers). What we do and how we do our jobs will change.
At least initially, some ways AI can be used in qualitative research are listed below (some of these are currently being done by machine):
- Follow up and ask probes of open-ended questions
- Summarize open-ended questions
- Create initial discussion guide
- Create summations and provide input into reporting
- Provide transcriptions and translations
- Scrape data from multiple sources, analyze such data, and suggest questions, key findings, etc.
One challenge will be to understand what training will be needed for qualitative researchers to continue to be successful. One of the early challenges for market research and for organizations like QRCA and qualitative research is understanding what will change in a way that allows us to create new codes and standards that provide our clients with the confidence in our work products while protecting research participants and our members from irregular and even abusive uses of this technology. The Insights Association and ESOMAR are both working on such codes. QRCA will look to see how our Code of Conduct will be affected by AI and suggest appropriate changes. Please feel free to weigh in with your thoughts and suggestions on this topic by emailing me at michael@nicholsresearch.com.
Best Regards,
Michael Mermelstein
