Artificial intelligence (AI) bots generated 76% of responses to a UK government consultation on tobacco and vaping restrictions in October 2023, with almost all opposing a ban on disposable vapes, researchers have found.

Of 118,756 submissions, 90,835 were machine-generated, according to researchers from the universities of Edinburgh and Bath. Their findings were published in the journal Tobacco Control.

The researchers described the incident as what appears to be the first documented case of large-scale AI interference in a formal public health policy consultation. The Department of Health and Social Care (DHSC) identified and excluded 90,835 submissions classified as fraudulent before assessing public opinion.

Officials Detected And Removed Thousands Of Fake Responses

Officials used automated bot-detection screening after finding that the entries were sent at high frequency from a small cluster of internet protocol (IP) addresses. The submissions also contained heavily duplicated wording.

All but two of the machine-generated responses selected the same multiple-choice option opposing restrictions on disposable vapes. Virtually all claimed to have learned about the consultation through social media. The screening prevented the campaign from distorting the government’s assessment of public opinion.

Once the machine-generated entries were excluded, 79.6% of valid responses supported restrictions on disposable vapes. The consultation preceded the UK’s ban on the sale or supply of single-use vapes, which took effect in June 2025.

The Tobacco and Vapes Act received Royal Assent in April 2026. The findings suggest the volume of submissions alone could have presented a misleading picture of public opposition had the automated entries not been detected.

Researchers Highlight Possible Tobacco Industry Links

The researchers said they could not definitively establish the campaign’s commercial origins. The researchers compared the operation with previous efforts by transnational tobacco companies to influence public health consultations, including mass recruitment drives and postcard campaigns. They did not establish who was behind this operation.

Action on Smoking and Health (ASH) described the use of AI bots to overwhelm health consultations as ‘deeply concerning’. The organisation said public health decisions should be based on sound evidence and expertise, rather than manipulated response volumes.

AI Could Make Future Campaigns Harder To Detect

Experts and think tanks warned that rapid advances in AI could make similar campaigns more difficult to identify. Automated submissions can create the appearance of widespread public support for a position, even when the responses do not reflect genuine public opinion.

Researchers also warned that consultations run by local councils and other bodies with limited technical resources could be particularly vulnerable. Without specialist screening systems, covert manipulation may go unnoticed by both officials and the public.

Public consultations allow citizens to participate in government decision-making, but stronger verification measures could make it harder for genuine respondents to take part. Weak safeguards, meanwhile, leave consultations open to mass manipulation, potentially undermining public trust.

Analysts called for bot-detection tools to be strengthened as automated campaigns become more sophisticated, particularly in settings with limited resources.

Think tank researchers advocated combining advanced screening methods with greater transparency about how consultation responses are processed. They cautioned that no single verification check could guarantee protection against manipulation.

Source link