Artificial intelligence has sparked an uncomfortable question across the research industry: if synthetic data becomes sophisticated enough, what happens to traditional market research?
It's a question that many researchers, technology providers and client-side insights teams are actively exploring. Yet framing the debate as a choice between human and synthetic data risks missing the bigger opportunity.
At Esomar Congress 2026, Wilko Rozema, SVP, EMEA MRA Sales at Dynata, explored a different perspective in his presentation, "Human Truth, Synthetic Intelligence: How Synthetic Data Can Enhance Research Without Replacing People": not whether synthetic data will replace human research, but how it can enhance it.
As synthetic data gains momentum, conversations often focus on its benefits. Faster project timelines. Greater efficiency. Improved feasibility where hard-to-reach audiences are involved. Use cases such as audience augmentation, digital twin modelling and survey feasibility assessment demonstrate the potential. The opportunity is significant.
However, speed alone does not create better decisions.
The real challenge facing researchers today is trust.
Synthetic models can generate highly convincing outputs, but convincing doesn't always mean accurate. AI systems can hallucinate, amplify bias, inherit limitations from training data and create outputs that appear credible while failing to reflect reality.
For an industry built on helping organisations make confident decisions, this presents a critical challenge.
The question is no longer whether synthetic data can be used. The question is when it should be used, how it should be validated and where human data remains indispensable.
Synthetic Data Needs Ground Truth
The effectiveness of any synthetic model depends on the quality of the source data used to build it.
Without robust human-derived data, synthetic outputs become increasingly detached from the behaviours, attitudes and contexts they aim to represent.
This is particularly important as organisations seek to apply synthetic approaches to increasingly complex business questions.
While synthetic data can deliver meaningful value in areas such as audience augmentation, survey feasibility assessment and digital twin modelling, measuring real-world attitudes and behaviours remains a more challenging application.
Put simply: synthetic data can extend our knowledge. It cannot establish truth on its own.
Validation Must Become the Industry Standard
One of the most important shifts required as synthetic research matures is moving away from evaluating outputs based on how realistic they appear.
Instead, synthetic datasets should be judged on their ability to reproduce meaningful research outcomes.
This requires rigorous validation processes that assess fidelity, utility, privacy protection and overall decision-making confidence.
Validation cannot be treated as an optional step performed after data generation. It must be embedded throughout the process.
As researchers, our responsibility is not simply to generate more data. It is to generate evidence that decision makers can trust.
The Rise of Hybrid Research
The future of market research is unlikely to be purely human or purely synthetic.
Instead, a hybrid model is emerging.
Human research continues to provide context, nuance, lived experience and behavioural truth. Synthetic methods offer scalability, speed and the ability to explore scenarios that may otherwise be difficult or costly to investigate.
Together, they can create a more powerful framework for generating insights than either could deliver alone.
The industry's role is not to choose between humans and machines. It is to understand how each contributes to better evidence, stronger decisions and greater confidence.
As synthetic methodologies continue to evolve, one principle should remain constant: innovation must enhance trust, not replace it.
Because in research, truth still starts with people.
Watch Dynata's Esomar Congress presentation here.
