AI bots flooded UK smoking ban consultation with fake opposition
Researchers found 90,000 automated responses trying to block vape restrictions, though officials removed them before finalizing rules.
By Charlotte Dubois ·
University of Edinburgh researchers confirmed on Tuesday that automated software generated 76% of the submissions to a major government consultation on tobacco control. The study, published in the journal Tobacco Control, identified 90,000 fake responses out of 118,756 total entries regarding the implementation of the Tobacco and Vapes Bill.
The UK government launched the public consultation in October 2023 to gather feedback on how to enforce a lifetime smoking ban. The legislation, which became law in April 2026, prohibits the sale of tobacco to anyone born in 2009 or later. The consultation also covered new restrictions on electronic cigarettes, including a ban on selling vapes to under-17s and a prohibition on sweet flavors marketed toward children.
Dr. Nason Maani, a senior lecturer in inequality at the University of Edinburgh and the lead author of the study, explained that the automated responses were not typical AI-generated text. They did not contain excessive punctuation or unusual vocabulary. Instead, the Department of Health and Social Care data showed a pattern of duplicate multiple-choice answers sent from a small number of IP addresses at high frequency.
The bots primarily targeted the vape restrictions. Dr. Maani noted that the raw data, including the automated entries, would have shown that 81% of respondents opposed limiting disposable vapes. Among the genuine human respondents, however, roughly four in five supported the proposed restrictions. The coordinated nature of the bot campaign mirrored tactics historically used by tobacco lobbyists to create an impression of overwhelming public opinion.
The Department of Health and Social Care stated that its control measures detected and removed the fake responses. The department said the clean data helped inform the final regulations for the smoke-free generation initiative. Officials confirmed they would remain alert to similar interference in future consultation processes.
Dr. Maani observed that this tactic is consistent with known strategies for rigging public opinion. While the specific actors behind this campaign are unknown, the pattern matches the history of tobacco industry interference. The UK scored 45 out of 100 on the Tobacco Industry Interference Index last year, indicating a low level of company meddling. Researchers and the non-profit news outlet The Examination have previously found attempts by cigarette giants to sway the bill.
The discovery raises concerns for other nations that may lack the detection tools used by the UK. Dr. Maani noted that while this specific attempt was caught, the dynamic of automated interference is likely to continue. He emphasized the importance of deliberative approaches and live participation methods that are harder to automate. The final rules now reflect the views of genuine respondents rather than the inflated numbers from the bot campaign.
