Questions to consider around AI and market research

Matthew Gullett, Bellomy’s SVP of Insights Technology and a driving force behind Bellomy AI Analytics, has thoughts about the future of market research in a world with AI.  

A couple questions about AI and market research that deserve our consideration: Does AI output shape people’s opinions? And does (or will) AI-generated output alter how we communicate?

On the surface, I think it’s clear that the answer to both questions is yes. From the perspective of market researchers, having awareness around these questions along with appropriate talking points and methodological strategies may prove important. So let’s dive in a bit deeper.

Q1: Does AI output shape opinions?

The short answer is, I certainly hope so. AI is being used to produce blog posts, news articles, video transcripts, movie clips and much more. These communication vehicles have always been used by marketers, politicians and others to get attention, drive awareness and shape consumer opinions and perspectives. AI output should contribute to this as much as prior human-only outputs.

There is a good deal of controversy around the use of AI, as seen recently in the Writers Guild strike. Many question the authenticity/veracity/quality of AI-written articles. We’ve seen news stories about poorly written articles that contain factual errors. However, there are potentially hundreds of thousands of AI-written articles on blogs and other news media that are accurate and informative.

Considering all of this, an additional question arises: Does output written by AI shape opinions in ways we need to be watchful of? I think the answer here also is yes. Some have pointed out the propensity of AI to take moderate positions on most debatable topics while leaning into topics related to politics. As consumers become more inundated with AI-produced content, the market research community needs to understand how the content might shape research objectives and outcomes. We also need to be on alert for scammers and those who use media to intentionally manipulate by pushing large-scale, mass misinformation campaigns using AI to get through media controls. We also need to be on alert for scammers and those intent on spreading large-scale misinformation campaigns using AI to circumvent media controls.

Q2: Does/will AI-generated output alter how we communicate?

I think the answer to this second question — which I find as interesting at the first — is again, I hope so. We’ve seen rapid change in human communication in the past 40 years. Consider things like texting, emojis, tweets and the 800-word-count blog post targets. As people embrace AI to augment and replace human writing in emails, blogs and communication, we would expect some changes to take place around perceptions on grammar rules, information flow through articles and more. I think changes like this will be positive overall. For example, my wife teaches sixth-grade English and will likely welcome improvements in these areas over time. But what does this mean for market research? We can’t predict the future, but I have some top-level thoughts.

• Survey verbatims are likely to be augmented by AI in the future. Browsers will likely embed generative AI directly into the browser experience, much like spellcheck and Grammerly.  Survey respondents will have, at their fingertips, easy ways to use generative AI to fill in form fields.  On a related note, respondents might be able to use browsers/browser extensions to summarize a survey’s long text blocks (e.g. instructions and introductions).
   
• We need to keep pace with AI-influenced changes in grammar and structure. How we present information in surveys and other research instruments has always needed to follow good grammar and structure, and as these evolve with AI, we will need to evolve alongside it.
   
• AI is likely to change the mix of knowledge people have on any topic. It is very likely that people will use AI to answer many everyday questions, and it follows that AI’s capacity to synthesize information across sources is likely to change the mix of knowledge people have on any topic. Whereas previously knowledge has been acquired through a variety of sources (books, Google searches, websites, etc.) AI will present people with synthesized summaries.
   
• We need a response to survey-takers using AI to fill in knowledge gaps related to questions they’re being asked. Measures might include an AI-related code of conduct for survey takers, adopting survey collection methodologies (like video) that address this concern, or even embracing respondents using AI to enhance their knowledge while taking surveys.
   

Our goal is to responsibly maximize AI's benefit to our clients

Bellomy’s team of researchers and technologists are working in lock-step to chart a course that enables positive AI-enriched research experiences and outcomes. From streamlined research workflows to AI-enriched respondent experiences and defensive tools to monitor, detect and address suspicious activity, Bellomy is equipped to serve clients with the right mix of traditional and leading-edge technology and techniques.

Written by Matthew Gullett, Bellomy’s SVP of Insights Technology. An employee of more than 20 years, Matt loves thinking and writing about AI and is a driving force behind Bellomy AI Analytics.