**
The landscape of opinion polling is undergoing a significant transformation, driven by the rapid advancement of artificial intelligence. A French company, Naratis, is pioneering this shift by utilising AI to conduct political opinion polls more efficiently and accurately than traditional methods. Founded by 28-year-old engineer Pierre Fontaine in 2025, Naratis aims to revolutionise qualitative research, a sector known for its slow and costly processes. As the polling industry grapples with declining response rates and public scepticism, AI presents a promising solution that could redefine how opinions are measured and understood.
The Rise of AI in Political Polling
Naratis is not alone in its endeavours; several start-ups in the US, such as Outset, Listen Labs, and Hey Marvin, are also leveraging AI for polling in commercial sectors. However, Naratis claims to be the first to apply this technology specifically to political opinion polling. By integrating conversational AI into their methodology, the company seeks to uncover not just what people think but how they formulate their opinions and how these perspectives evolve over time.
Traditional qualitative studies often involve lengthy interviews with small groups or individuals, requiring significant time and resources. Naratis has streamlined this process, allowing for rapid data collection and analysis. According to Fontaine, their approach is “10 times faster, 10 times cheaper, and 90% as accurate as human polling.” This speed is achieved through what he describes as “parallelisation,” where multiple AI agents conduct interviews simultaneously, vastly improving efficiency.
Addressing the Challenges of Modern Polling
The introduction of AI into polling comes at a critical juncture for the industry. Response rates have plummeted over the years, dropping from over 30% in the 1990s to below 5% today, as noted by AI consultant Stéphane Le Brun. This decline poses challenges for the accuracy and representativeness of polling data, contributing to a growing distrust among the public.
Critics of AI polling have raised concerns about its reliability, particularly in light of past polling failures such as the 2016 election outcomes in the US and Brexit. However, Fontaine argues that these issues are primarily associated with quantitative polling, which focuses on predicting outcomes rather than understanding the nuances of public sentiment. Qualitative research, he posits, is more about grasping the complexities of opinions and reactions to political messaging.
Established polling firms are also adapting to the AI trend. Ipsos, for instance, employs AI in market research to analyse video footage of respondents rather than relying solely on self-reported data. This shift allows researchers to observe behaviours directly, enhancing the depth of their insights.
The Ethical Implications of AI-Driven Polling
Despite the advantages, the use of AI in polling raises ethical concerns. In politically sensitive contexts, firms like Ipsos are cautious about employing AI-generated respondents in political surveys, with OpinionWay’s CEO, Bruno Jeanbart, stating they would never publish polls based on AI data due to trust issues. The possibility of AI systems “hallucinating,” or fabricating plausible but inaccurate answers, further complicates the trust factor in polling data.
Moreover, synthetic data—responses generated rather than collected from actual individuals—raises fundamental questions about the validity of the information being measured. As AI continues to enhance polling capabilities, the balance between augmenting human data and generating synthetic responses will be critical, particularly in the political arena.
The Future of Opinion Polling
The future of opinion polling is likely to be a hybrid model, where AI enhances traditional methods rather than completely replacing human input. Naratis is at the forefront of this evolution, focusing on transforming surveys into dynamic conversations. As they expand their reach, the company aims to gather insights at unprecedented scales while ensuring the integrity of the data collected.
The impact of AI on polling will hinge on how the technology is deployed, communicated, and regulated. With economic pressures driving the industry towards greater automation, the challenge will be to restore public trust in polling practices. Whether AI will be a tool for greater transparency or a source of further scepticism remains to be seen.
Why it Matters
The integration of AI into opinion polling signifies a pivotal moment in how public sentiment is gauged and understood. As traditional methods face challenges from declining response rates and increasing public distrust, AI offers a pathway to more efficient, cost-effective, and insightful polling. However, the ethical implications and the importance of human oversight cannot be overlooked. Ensuring that the technology is used responsibly and transparently will be essential in shaping the future landscape of public opinion research. The way forward will depend on not only the technological advancements but also the collective commitment to uphold the integrity of data collection and analysis in an increasingly automated world.