What role could AI play in market research surveys?

The world is racing to understand the impacts, both positive and negative, of artificial intelligence (AI). As market researchers, we share the responsibility to help determine how these tools can be used – where it’s adding value or simply where it’s a case of AI-washing ​​(Kuzmina, 2023), claiming to use AI when in reality, there is little to no interaction with it (Rosencrance, 2023). As the technology continues to advance, we must understand the areas in which AI could be used to disrupt the market research industry. Below are some frequently asked questions from our customers and ideas on how to proceed with caution.

Can AI mimic humans to generate open-ended responses in surveys?

Technically, the answer is yes. However, the probability of this happening frequently is unlikely. There is a certain flow between a survey and its respondent. When considering the amount of time required to answer a question, it is safe to assume that it would take longer for a respondent to leave the survey, engage AI and copy/paste the generated response into the survey, than it would for the respondent to simply write the answer themselves. Thus, this should not be an overwhelming concern to the industry simply taking into account the loss of time and increased energy required of the respondent to utilize a tool such as ChatGPT in a survey environment.

Can AI create surveys without experience or human intervention?

Of course. However, a major difference between a human-created survey and an AI-created one is the incorporation of industry experience. AI can generate a solid survey that may serve as a starting point for a market researcher to meld into a comprehensive, well-designed questionnaire, but one must be alert to the fact that AI cannot accurately account for considerations such as implicit bias, methodological errors, survey structure and inclusion best practices. If one decides to engage AI in the survey building process, be sure to review the results, budget time for editing and confirm no confidential information has been accidentally included (Kharpal, 2023).

AI can generate survey responses, but can it generate them at scale?

At this time, AI does not have the ability to mass produce fraudulent survey responses; however, this is a growing concern as the technology advances. As a leader in the market research space, Dynata uses machine learning (ML) to detect and eliminate fraud by implementing authenticity checks on responses, incorporating increased respondent validation, building barriers to prevent reward redemption by bad actors and identifying behavioral anomalies associated with IP addresses. While AI continues to develop, we see opportunity in building processes and governance to grow with it (Frank, Greene, 2023).

In the end, AI does not currently have the capability to completely replace human researchers or to effectively collect data, and large language models are unreliable when predicting people’s responses to text-based survey questions. Still, just as corporations are working to protect their data and governments to protect their citizens (Satariano, 2023), market researchers must work to protect the methods in which their surveys are created, distributed and interpreted. The solution is ever-changing but Dynata’s approach is to continuously track developments in AI and ML to safeguard and supplement its current processes in a strategic and steady manner.

Sources

  • Kuzmina, A. (2023, June 6). Digital Snake Oil or Genuine AI? How to avoid AI-washing in Market Research. GreenBook | Market Research. Retrieved June 15, 2023, from https://www.greenbook.org/mr/the-prompt/digital-snake-oil-or-genuine-ai-how-to-avoid-ai-washing-in-market-research/

  • Rosencrance, L. (2023, June 12). AI Washing: Everything You Need to Know. Techopedia. Retrieved July 3, 2023, from https://www.techopedia.com/ai-washing-everything-you-need-to-know/2/34841#What_is_AI_Washing

  • Kharpal, A. (2023, May 2). Samsung bans use of AI-like ChatGPT for employees after misuse of the chatbot. NBC News. Retrieved June 15, 2023, from https://www.nbcnews.com/tech/tech-news/samsung-bans-use-chatgpt-employees-misuse-chatbot-rcna82407

  • Frank, A., & Greene, N. (2023). Beyond the Hype: The Impact of Generative AI on Marketing. Gartner. Retrieved June 15, 2023, from https://www.gartner.com/en/webinar/487667/1143179

  • Satariano, A. (2023, June 14). E.U. Takes Major Step Toward Regulating A.I. The New York Times. Retrieved June 15, 2023, from https://www.nytimes.com/2023/06/14/technology/europe-ai-regulation.html