Stakeholder Perspectives on the Revision of the EU Tobacco Products Directive: A Census-Level Thematic Analysis of a Public Consultation using a Large Language Model
This study utilizes a large language model to conduct a census-level thematic analysis of 682 technical submissions to the EU's 2026 tobacco product revision consultation, revealing a predominant stakeholder consensus favoring regulatory flexibility and tobacco harm reduction that challenges the prevailing political narrative of precautionary restriction.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the European Union, the rules governing tobacco and nicotine products are set by a major law known as the Tobacco Products Directive. This legislation dictates how these items are made, sold, and advertised, with the ultimate goal of reducing the number of people who use them. Recently, the European Commission launched a massive public inquiry to decide whether these rules should be tightened or changed. The process was designed to gather evidence from everyone, from ordinary citizens to large corporations and health organizations. A central tension in this debate involves the concept of "harm reduction." This idea suggests that while nicotine is addictive, the primary danger comes from burning tobacco. Therefore, switching to non-combustible alternatives, such as electronic cigarettes or heated tobacco products, could significantly lower health risks for adult smokers. However, others argue that any new nicotine product poses too many unknown risks, particularly for young people, and that the safest path is to restrict these products as strictly as possible. Understanding which side holds more weight in the technical evidence submitted by experts is crucial for shaping future laws that affect millions of lives.
To navigate the overwhelming volume of feedback received during this inquiry, a researcher named Konstantinos Farsalinos turned to a powerful new tool: a large language model. This is a type of artificial intelligence capable of reading and understanding vast amounts of text, far beyond what a human team could process in a reasonable time. The European Commission had received over 82,000 responses, but most were short comments limited to a few thousand characters. The most detailed arguments, however, were hidden in hundreds of attached documents, often lengthy technical reports in PDF format. Farsalinos focused his study on 682 of these substantial submissions. He used the artificial intelligence to read every single document, identify the main topics discussed, and determine the stance of each author regarding harm reduction and the strictness of future regulations. This approach allowed him to map the landscape of expert opinion with a level of detail that was previously impossible.
The analysis revealed a clear and surprising divide in how different groups viewed the evidence. The most common theme across all submissions was a call for decisions to be based on solid scientific evidence and rigorous data. Nearly three-quarters of the documents emphasized this need. Closely following this was a strong argument for "harm reduction" and regulations that match the level of risk a product poses. About two-thirds of the submissions supported the idea that rules should be flexible enough to allow for safer alternatives to traditional cigarettes. When the researcher looked at who was supporting these views, a distinct pattern emerged. While non-governmental organizations and public authorities were more likely to oppose harm reduction and favor stricter rules, the majority of other groups took the opposite view. Consumer organizations, academic institutions, and individual citizens, including many who identified themselves as scientists and healthcare professionals, overwhelmingly supported a flexible approach that recognizes the potential benefits of switching to less harmful products. In fact, nearly half of all the detailed submissions were in favor of harm reduction, while only a small fraction opposed it.
The study also highlighted how different stakeholders interpreted the same facts in different ways. For instance, almost everyone agreed that scientific evidence was important, but they used it to support opposite conclusions. Those favoring harm reduction used the evidence to show that non-combustible products are significantly safer than smoking. Those favoring strict restrictions used the same pillar of evidence to highlight gaps in long-term data and concerns about youth usage. Similarly, the issue of "illicit trade," or the illegal sale of tobacco products, was a major topic. Industry groups often argued that strict rules would drive the market underground, while public health groups sometimes viewed these arguments as a tactic to avoid regulation. The research showed that the debate is not simply a clash between industry and health advocates, but a complex disagreement between two coherent ways of looking at the evidence. One side prioritizes the immediate protection of youth through prohibition, while the other prioritizes the long-term health of adult smokers through risk-proportionate choices.
One of the most significant findings was the role of the "citizen" in this high-stakes policy debate. The study found that many individuals who submitted comments as private citizens were actually doctors, researchers, and scientists. These experts, who often bypassed traditional institutional channels to speak directly to the government, were almost universally in favor of harm reduction and flexible regulations. This suggests that the voice of the scientific community is more supportive of harm reduction than the public narrative often implies. The researcher validated the accuracy of the artificial intelligence by manually checking a sample of the results, confirming that the machine's analysis was highly reliable. The study concludes that the technical evidence submitted to the European Commission does not support a single, unified demand for stricter rules. Instead, it reveals a substantial and well-argued case for a regulatory framework that balances the protection of young people with the needs of adult smokers seeking safer alternatives. Ignoring this evidence, the researcher suggests, would mean overlooking a vast amount of expert opinion that could shape more effective and life-saving policies.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.