A study in Tobacco Control found that 90,835 of 118,756 responses to the UK government's 2023 tobacco and vapes consultation were identified as automated, almost all opposing restrictions on disposable vapes.
Three-quarters of the responses to the UK government's 2023 consultation on creating a smoke-free generation and restricting disposable vape sales were generated by bots, according to an analysis published in the journal Tobacco Control. The Department of Health and Social Care identified 90,835 of the 118,756 responses, or 76 percent, as fraudulent using bot-detection methods.
Almost all of the automated responses targeted a single multiple-choice question, opposing proposed restrictions on disposable vape sales. The submissions were often duplicated many times, sent at high frequency from a small number of internet addresses, and reported social media as the source through which respondents had learned of the consultation. All but two of the 90,835 automated responses opposed the restrictions.
Removing them changed the picture completely. After the bot responses were excluded, 77 percent of valid respondents supported restrictions on disposable vape sales, the opposite of what the unfiltered data suggested. The department excluded the automated submissions from its decision-making, and the government later passed the tobacco and vapes legislation including the ban on businesses selling or supplying disposable vapes.
The study, by Nason Maani of the University of Edinburgh and Anna Gilmore of the University of Bath, draws on published consultation reports and information obtained through a freedom-of-information request to the department. Its claim that this appears to be the first documented instance of large-scale automated interference in a formal public health consultation is the researchers' qualified wording, not proof that no earlier attempt occurred, and the researchers did not identify a particular AI model behind the submissions.
The stakes are about how governments count public opinion. Researcher Nason Maani warned that future campaigns may become harder to catch, saying: "the threat is that future campaigns become harder to catch, and what we can't catch, we don't know about." A campaign that makes a narrow position look like a mass movement does not need to succeed to damage trust in the process; it only needs to leave doubt about whose voices were counted.
The unresolved design problem is the trade-off the researchers identify: tighter screening can deter genuine participants, while weak screening risks distorting the record, and every government that asks for comments online now faces the same choice in its next consultation.