🔮 Synthetic users that think like our customers? AI analyzing 20 hours of interviews in minutes? This isn't sci-fi—it's happening NOW in UX research.
I recently dove into this interesting HBR article on how generative AI is transforming market research, and it got me thinking deeply about the implications for UX and Product Research. Existentialism anyone?
Korst, Puntoni, and Toubia (2025) argue that generative AI is poised to revolutionize how companies conduct market research–and is already starting to. AI will undoubtedly transform how we, within the research community, plan, design, conduct, analyze, and socialize our research. And I find myself mostly curious and excited by the possibilities…while simultaneously being so nervous I could puke.
🚀 The Opportunity
Gen AI is already being used to synthesize data, analyze transcripts, and even generate synthetic users to help us understand those hard-to-reach audiences (think B2B or sensitive research topics).
Early adopters are reporting significant productivity gains (62% using it to synthesize lengthy interview transcripts, 58% for data analysis!). I can definitely attest to this. I recently used AI to help synthesize 20 hrs of interviews--what would have been at least 80 hrs of coding and synthesizing only took 14 hrs.
The idea of "digital twins" for testing concepts and refining pitches is intriguing, though it raises some serious ethical, privacy, and security questions...
🧐 This technology comes with some critical caveats:
Bias, bias, bias: 77% of researchers are concerned, and rightly so. As the HBR article points out, these models can perpetuate existing societal biases (Santurkar et al., 2023). We, as researchers, must be the guardians against that.
Synthetic users aren't real users: They lack the nuance, emotional depth, and unpredictable variance of actual human beings. One study found their responses were less varied and sensitive to question-wording (Bisbee et al., 2024).
The "great" data gap: Only 31% of researchers rate the value of Gen AI-generated data as "great." Accuracy and reliability are still a concern.
So, where does this leave us?
I see a few key takeaways for the UXR community:
1. Augment & assist, don't replace: use AI to handle the mundane while we focus on deeper human insights. Gen AI can augment our workflows, but our skills in critical thinking, qualitative analysis, ethnography, social and cultural norms, body language, and ethical research practices are more crucial than ever.
2. Become bias detectives: Develop frameworks to identify and mitigate algorithmic bias. We need to be at the forefront of identifying and mitigating bias in these models. IMHO, this is the new frontier and a critical requirement for future researchers.
3. Champion human-centeredness: In a world of synthetic data, we must advocate even more strongly for the importance of real user interaction and deep, contextual understanding. I'm endlessly curious about where this is all headed.
What are your thoughts? Have you used AI in your research process yet? What opportunities or concerns keep you up at night?
👇 Share your experiences below.
#AIinUXResearch #ResearchEthics #FutureofResearch #UXResearch
References
Bisbee, J., Clinton, J. D., Dorff, C., Kenkel, B., & Larson, J. M. (2024). Synthetic replacements for human survey data? The perils of large language models. Political Analysis, 32(4), 401-416. https://doi.org/10.1017/pan.2024.5
Korst, J., Puntoni, S., & Toubia, O. (2025). How Gen AI is transforming market research. Harvard Business Review, 103(3), 86-95.
Santurkar, S., Durmus, E., Ladhak, F., Lee, C., Liang, P., & Hashimoto, T. (2023). Whose opinions do language models reflect? In Proceedings of the 40th International Conference on Machine Learning (pp. 29971-30004).
—
Header Image Description: A vibrant, surreal collage on a cream background depicts a UX researcher interviewing a synthetic user. The researcher, a photographic collage of a professional woman from the shoulders up, holds a tablet and looks attentively at the synthetic user. The synthetic user, the central element, is a chaotic yet harmonious collage of various photographic robot parts, digital interface fragments (glowing lines, distorted buttons), and organic textures (colorful paint splatters, brushstrokes in blue, green, purple, and orange). Abstract geometric shapes are layered throughout, occasionally overlapping the figures.