Research has found that artificial intelligence tools were used to generate the bulk of responses to a 2023 UK government consultation on a proposed ban on disposable vapes, in what appears to be a coordinated effort to influence health policy.
According to the study, roughly three-quarters of the submissions to the consultation were machine-generated, amounting to tens of thousands of fake or automated responses. The activity targeted the government’s tobacco and vapes bill, which included measures to restrict the sale of single-use vaping products.
Scale and suspected coordination
Researchers described the pattern of replies as consistent with coordinated use of AI to produce mass responses, marking what is thought to be the first organised attempt to deploy generative tools to disrupt UK health policymaking. The volume and uniformity of the entries raised questions about the integrity of the consultation process and the ability of existing systems to detect inauthentic participation.
Outcome and official process
Despite the flood of automated submissions, the research indicates the campaign did not succeed in stopping the bill from passing into law. The consultation remained part of the formal legislative process, and the episode did not ultimately prevent the policy from advancing.
Implications for public consultations
The incident has prompted concern among analysts about the vulnerability of public engagement exercises to manipulation by AI. It highlights the need for government departments to reassess verification and filtering mechanisms used in consultations to ensure that responses reflect genuine public opinion rather than mass-produced content.
As generative AI becomes more accessible, researchers warn that similar tactics could be used to influence a range of policy areas. The episode underscores growing calls for improved safeguards and transparency in how governments collect and evaluate input from the public.