by Danny Sheahan - This is the ChatGPT paraphrased version of the original article here. I would prefer if you read the original article and refer to this one purely out of curiousity. I have no idea how ChatGPT would feel about this.

11 April 2023

Almost thirty years ago, I created, managed, and analyzed my first paid customer satisfaction survey. To optimize response rates, we used physical mailers with printed questionnaires and pre-paid return envelopes. Since then, the survey research landscape has transformed significantly, with the most substantial shift being the transition from paper-based surveys to online distribution. Although paper and telephone surveys still have their place, online surveys are now the prevailing method, thanks to the rise of the internet and widespread smartphone adoption.

We now seem to be on the cusp of another revolution, as Generative Artificial Intelligence (AI) starts to influence not just market research but all aspects of our professional and personal lives. Like many others, I have been experimenting with ChatGPT and similar Generative AI applications to understand their potential impact on market research.

This revolutionary technology brings both opportunities and challenges, some of which are evident now, while others will become clearer in the coming months and years. For the market research field, I see three immediate opportunities and three potential risks:

Opportunity #1 - Faster data exploration and reporting: ChatGPT's summarization capabilities are impressive, and it's likely that visual summaries of quantitative data with natural language text prompts will soon follow. Access to ChatGPT's API and other APIs will expedite this process.

Opportunity #2 - Quicker insights: Human expertise is still required for generating insights, but AI can help us reach that stage more rapidly by completing routine tasks in the market research survey workflow.

Opportunity #3 - Teaching AI to produce insights: While this is more challenging, it can be achieved with sufficient data. Although true innovation might be limited since Generative AI relies on existing information, creative extrapolation can inspire researchers to explore data they may not have considered before.

Risk #1 - Poorly designed questionnaires: Initial attempts at using ChatGPT to design survey questionnaires have resulted in suboptimal designs. This is expected to improve over time, but for now, it could lead to an increase in poorly designed questionnaires, which in turn results in risk #2.

Risk #2 - Erroneous analysis and conclusions: Faulty conclusions may lead to disastrous business decisions. At Delve Research, we are testing historical data on generative AI systems to gauge their accuracy. While analysis speeds are impressive, some conclusions have been less so. The hallucination effect should be approached with caution until improvements are made.

Risk #3 - Enhancing survey fraud through smarter bots: Survey fraud is already a significant issue for the market research industry, and Generative AI may exacerbate this problem by enabling bad actors to create fraudulent respondent bots. The industry must respond swiftly and firmly to maintain its credibility.

The rapid evolution of this technology is both exciting and challenging. At Delve Research, we are investing in and adapting to these developments. If you would like to share your thoughts on this subject or learn more about our use of generative AI, feel free to email me directly at danny@delve-research.com.

A few notes:

This article was paraphrased by ChatGPT v4.0 from the original text at https://www.delve-research.com/generativeaimarketresearch/

The image for this article was generated by Midjourney using my prompts

© Delve Research 2023, paraphrasing notwithstanding