Stories

Thu, 30 Apr 2026

Thu, 30 Apr 2026 Will AI lead to more accurate opinion polls?

It's cheaper and faster to collect people's opinions using AI, but will it make polls more accurate?
A French company called Naratis has developed an AI-powered polling system that can conduct conversations with people and analyze their opinions in real-time, providing more accurate results than traditional human-led polls. The system uses a combination of three AI agents to process responses, ensuring accuracy and authenticity. Founded by engineer Pierre Fontaine in 2025, Naratis aims to automate the labor-intensive aspect of opinion research and provide deeper insights into public opinion. According to Fontaine, the AI-powered method is "10 times faster, 10 times cheaper and 90% as accurate as human polling". The system can gather responses from thousands of people within a day or two, allowing for rapid reaction to events. Naratis claims that its method replaces traditional qualitative research methods that involve weeks-long interviews with paid respondents. Critics argue that AI-powered polls have limitations, including the potential for "hallucinations" (inventing plausible but incorrect answers) and producing "common sense" responses shaped by societal norms rather than individual opinions. Additionally, there are concerns about trust, as people may be more candid in conversations with machines rather than humans. Established polling firms like Ipsos are also incorporating AI into their research methods, using digital twins (virtual models of real individuals) and synthetic data to analyze behavior and reactions. However, these techniques raise questions about the validity and interpretation of data generated by simulated respondents. The use of AI in polling is expected to continue, with companies pushing for greater automation due to economic pressures. However, advocates acknowledge that human oversight remains essential for validating results and taking responsibility. The industry's future likely lies in a hybrid model where humans work alongside AI systems to augment and validate data.