More than three-quarters of responses to a UK government consultation on disposable vapes were submitted by automated programmes, raising concerns about attempts to manipulate public health policymaking.

Research published in the journal Tobacco Control on 27 October found that bots submitted 90,835 of the 118,756 responses received during the government's 2023 consultation on tobacco and vaping legislation. The Department of Health and Social Care (DHSC) identified the submissions as fraudulent and excluded them from its decision-making process.

The findings were reported by The Guardian and published in Tobacco Control by researchers Dr Nason Maani of the University of Edinburgh and Professor Anna Gilmore of the University of Bath. The researchers described the incident as the first documented case of large-scale automated interference in a formal public health policy consultation.

Almost all the automated submissions opposed proposed restrictions on disposable vapes. According to the DHSC, many were duplicated repeatedly and submitted at high frequency from a small number of internet protocol (IP) addresses.

How Did the Fake Responses Affect the Results?

The scale of the activity could have significantly distorted the apparent balance of public opinion had the fraudulent submissions not been detected.

After the automated responses were excluded, 79.6% of valid respondents supported restrictions on disposable vape sales, according to the government's consultation findings. The result illustrates how a large volume of artificial submissions can make a particular position appear more widely held than it is among genuine respondents.

However, researchers have not established who organised the campaign or definitively identified the interests behind it. Although the pattern was consistent with coordinated interference, the findings do not prove that a particular company or industry group directed the activity.

Caroline Cerny, deputy chief executive of Action on Smoking and Health, told The Guardian that the use of AI bots to overwhelm a government consultation was deeply concerning. She warned that technological advances could be used to advance commercial interests at the expense of public health.

The DHSC said its safeguards had worked effectively because the fraudulent responses were detected and removed before they influenced the final decisions.

Read more

Why Was the Government Consulting on Disposable Vapes?

The 2023 consultation sought public views on several proposed measures, including creating a smoke-free generation, tackling youth vaping and restricting disposable vape sales.

Disposable vapes had become a concern because of their appeal to young people and their environmental impact. The devices contain batteries and other components that can be difficult to recycle, while discarded products contribute to electronic waste.

The government's consultation response recorded strong support for restrictions, including a proposal to prohibit the sale and supply of disposable vapes. The disposable vape ban was subsequently introduced under the Labour government led by Keir Starmer and came into force in June 2025.

Could AI Interference Affect Future Health Policies?

Researchers warned that future campaigns could be harder to detect as automated systems become more sophisticated. Public consultations help governments understand public concerns, but artificial submissions risk undermining their usefulness by making a coordinated minority appear to represent a much larger movement.

Azzurra Moores, associate director for information ecosystems at the think tank Demos, told The Guardian that interference could undermine public trust in consultation processes. She also noted that measures designed to block bots must not make it unnecessarily difficult for genuine members of the public to participate.

The researchers said stronger detection systems and safeguards would be essential to protect future consultations. The case highlights a growing challenge for policymakers: ensuring that public health decisions reflect authentic public input rather than manipulated response numbers.

Read more