Tens of thousands of automated responses were submitted to a British government consultation on tobacco and vapes, research published on Tuesday has found.
The Department of Health and Social Care used bot-detection methods to identify that 90,835 of the 118,756 responses it received to the 2023 "smoke free generation" consultation, or 76 per cent, came from bots.
The finding was disclosed in research published in the journal Tobacco Control by the health policy experts Dr Nason Maani, of the University of Edinburgh, and Prof Anna Gilmore, of the University of Bath, and reported by the Guardian.
All but two of those 90,835 automated responses answered the same question, in which they opposed the plan to ban disposable vapes. The submissions were "often duplicated many times and sent at high frequency from a small number of IP addresses", the DHSC said.
The episode "appears to be the first documented instance of large-scale automated interference in a formal public health policy consultation", the researchers said. The published research does not establish who sent the responses.
The consultation was run under Rishi Sunak's Conservative government. Keir Starmer's Labour government inherited the tobacco and vapes bill in 2024 and passed it into law, including the ban on businesses selling or supplying disposable vapes. The 90,835 automated responses were excluded from the department's decision-making before any of that happened.
Dr Maani said that while this attempt was identified, "the threat is that future campaigns become harder to catch, and what we can't catch, we don't know about. Such campaigns are a threat because they raise the possibility of a handful of powerful interests looking like a mass movement."
Caroline Cerny, the deputy chief executive of Action on Smoking and Health, said the evidence was "deeply concerning" and added: "Decisions that affect public health must be based on sound evidence and expertise, not manipulated response numbers."
Azzurra Moores, the associate director for information ecosystems at the thinktank Demos, said consultations were "one of the main ways government opens up its decision-making, so any interference with them is a serious concern". She added that the problem had no easy answer: "Every barrier that keeps bots out also makes it harder for real people to take part, and AI makes fake responses far harder to spot."
A DHSC spokesperson said: "We always have robust controls in place to prevent the manipulation of our public consultations, and those controls worked effectively in this case. As a result, the fake responses were removed, and the consultation helped inform the landmark action we are taking to create a smoke-free generation."
The arithmetic deserves a closer look, because it is remarkable in a way the coverage has passed over. Of the 90,835 automated responses, 90,833 answered a single question in a single direction.
That is an agreement rate of 99.998 per cent. No consultation answered by human beings has ever approached it. On any ordinary reading of consultation evidence, this is the clearest expression of public feeling ever recorded on the subject of disposable vapes, and it was held by nobody at all.
That leaves the two. Two of the automated responses did not answer that question. The research does not record what they did instead. They remain the only participants in the exercise whose position cannot be inferred from their IP address, and this newspaper considers them the most interesting respondents of 2023.
It is worth restating what a consultation is for. It is not a vote and never has been; it is an instrument for gauging the strength of feeling about a proposal. Strength of feeling has always been treated as something a person must supply. The 2023 exercise establishes that it can also be manufactured, duplicated and delivered at high frequency, which is the one direction the instrument was never calibrated to measure.
The Department of Unnecessary Research has reviewed the figures and finds that the consultation performed exactly as designed. It asked the public what it thought, and something answered.