A British public consultation from 2023 may have been flooded by bots. A new study has found signs of a large-scale automated campaign: more than 90,000 responses were classified as fraudulent, and almost all of them opposed restrictions on disposable e-cigarettes.
Researchers from the universities of Edinburgh and Bath analysed the case and published their findings in the journal Tobacco Control. They found signs of automated interference in a UK public consultation on measures to reduce tobacco and e-cigarette use among young people. The authors admit that this may be the first documented case of an operation of this kind, on a large scale, in a formal consultation on health policy.
More than 90,000 responses under suspicion
The consultation, held in October 2023, received 118,756 responses. Of these, 90,835, or 76.5%, were classified as fraudulent by the Department of Health and Social Care (DHSC), through its bot-detection mechanisms.
According to the study, almost all of the suspicious responses were limited to three multiple-choice questions. Of these, 90,833 expressed opposition to restrictions on the sale and supply of disposable vapes, and the same number said they had learned of the consultation through social media.
To reach these conclusions, researchers Nason Maani and Anna Gilmore analysed official reports and information obtained through freedom of information requests submitted to the DHSC. According to the UK government data cited in the study, the suspicious responses had characteristics inconsistent with a normal completion of the form: repeated submissions at a high frequency, coming from a small number of IP addresses.
The scale of this interference could have significantly changed how the results were read:
81.3% of participants would appear to oppose the restrictions, if all responses were considered legitimate;
but 79.6% of the participants considered valid supported the measures, once the responses classified as fraudulent were excluded.
It is important to note that these results do not necessarily represent the opinion of the British population as a whole.
Unknown perpetrators and limitations of the study
The scale of the operation and the uniformity of the responses suggest a coordinated campaign, but it was not possible to identify those responsible nor to determine whether commercial interests were involved. The study recalls that multinational tobacco companies have previously used mass mobilisation campaigns to influence public consultations.
The researchers also acknowledge that they could not independently verify the detection methods used by the DHSC, as they were not disclosed, and that they had no access to participants' metadata or demographic information. The study also does not show that the responses were produced using generative Artificial Intelligence.
Meanwhile, in the UK, the sale and supply of disposable e-cigarettes was banned on 1 June 2025. More recently, the Tobacco and Vapes Act received Royal Assent on 29 April 2026 and introduces new measures to control tobacco and vaping products, including the future ban on selling tobacco products to people born on or after 1 January 2009.
How to protect public consultations
For the study's authors, the evolution of bots requires stronger detection and protection mechanisms for public consultations. Among the solutions discussed are limiting the number of submissions per IP address, using CAPTCHA systems and creating mechanisms to identify potential conflicts of interest.
The study also warns that the risks may be greater in countries or institutions with fewer technological resources to detect automated campaigns.
