The pros and cons of AI-generated research respondents

Editor’s note: Lisa Boughton is the director of Angelfish Fieldwork and Angelfish Opinions. This is an edited version of an article that originally appeared under the title “Synthetic Respondents in Market Research: Risk or Reward?”

Market research is all about understanding real people – how they think, feel and behave. But what if those "people" weren’t people at all? Enter synthetic respondents: AI-generated participants designed to simulate human survey responses. Sounds futuristic, right? Some hail this as the next big thing in research, offering speed, scalability and cost savings. Others worry about authenticity, ethics and whether AI can ever truly replace human insight.

So, what’s the reality? Are synthetic respondents an exciting innovation or a potential risk to data integrity? Let’s dive in.

What are synthetic respondents?

Put simply, synthetic respondents are AI-generated personas that take part in surveys, focus groups and other research methods – without a real person behind them. These virtual participants are created using machine learning and behavioral modelling to mimic human responses, often based on vast datasets of real-world consumer behavior. 

The idea isn’t entirely new. AI-driven analytics and predictive modelling have been used in research for years. But instead of just analyzing existing data, synthetic respondents actively participate, providing answers just like a human respondent would. But can AI really replicate the complexity of human thought?

The benefits of synthetic respondents

There’s no denying that AI-powered respondents bring some big advantages to the table: 

Cost efficiency

There is no need to recruit, incentivize or manage real participants. AI can churn out responses at a fraction of the cost.

Scalability

Need 10,000 survey completions overnight? No problem. AI respondents don’t sleep. 

Flexibility

Synthetic samples can be tailored to any demographic, ensuring balance across age, income and behavior segments.

Speed

AI can analyze and respond instantly, slashing research turnaround times.

Bias reduction

In theory, AI removes human biases – no social desirability effects, no memory errors. But is that really the case? 

The risks and challenges of synthetic respondents

For all their potential, synthetic respondents come with big questions – and even bigger risks.

Lack of authenticity

AI models rely on historical data. But can they generate new insights, or are they just repackaging old ones? Market research is about exploring what’s changing, not just what’s already known.

Oversimplification 

Human decisions are complex, emotional and often irrational. AI, no matter how advanced, follows patterns and probabilities – it doesn’t feel. 

Ethical concerns

Who’s ensuring that synthetic respondents are being used responsibly? Without transparency, clients may not even realize that their data is AI-generated.

Regulatory and industry standards

The MRS Delphi Report on AI warns that AI in research should align with ethical standards, particularly in how data is collected and validated (registration required). The IQCS and MRS are actively working on AI guidelines, but the industry still lacks firm regulation. 

Trust and data integrity

Will businesses trust insights from AI-generated respondents? If synthetic samples become widespread, could we see data dilution, where real human behavior is no longer properly represented? AI can complement real research – but it’s not a replacement for genuine, human-driven insight.

Will AI and synthetic respondents replace human respondents?

Probably not. At least, not entirely. While synthetic respondents might work well for predictive modelling and early-stage concept testing, they struggle with:

  • Emotional and subconscious decision-making – Humans buy products based on gut feelings, nostalgia and social influence – things AI can’t fully replicate.
  • Qualitative research – A virtual respondent can’t engage in a genuine conversation, pick up on tone of voice or uncover those all-important “aha” moments.
  • Cultural and social contexts – Trends, slang and cultural attitudes shift fast. AI trained on past data might not keep up. 

What is the future of AI, synthetic respondents and human insight?

Rather than AI vs. humans, the future of market research will likely involve both. AI for large-scale data processing, predictive insights and synthetic modelling and humans for qualitative depth, emotion and real-world unpredictability.

Brands will need to decide when AI makes sense and when human respondents are irreplaceable. Our take? When you want insights that truly connect with real people, you need real respondents. No AI can replace human stories, emotions and experiences. So, while synthetic respondents might be an interesting tool, for high-quality, trustworthy research, humans still have the final say.