AI Bots and the Integrity of Digital Consultation: Lessons for UK Organisations

Research reported by The Guardian on 27 October 2026 has revealed that three-quarters of all responses to the UK government's 2023 consultation on the disposable vapes ban were machine-generated. The coordinated campaign used AI bots to submit tens of thousands of fabricated responses in an attempt to prevent the tobacco and vapes bill from passing into law. Researchers believe this to be the first confirmed coordinated use of AI to attempt to derail health policymaking in the UK. The effort ultimately failed to prevent the legislation, but the scale of the manipulation — with AI-generated content vastly outnumbering genuine human responses — raises important questions about the integrity of digital public engagement processes.

Why this matters for UK organisations

The immediate story concerns government policy, but the underlying technique has much broader relevance. The same AI-driven approach used to flood a public consultation could be applied to manipulate online reviews, customer feedback platforms, employee engagement surveys, market research exercises, or any other mechanism that aggregates text-based input at scale. For organisations that use consultation data, stakeholder feedback or customer sentiment to inform strategic decisions, the authenticity of that input can no longer be assumed without some form of verification. This is also relevant for organisations involved in regulatory affairs or public policy engagement: if consultation processes can be distorted at this scale, the signals that organisations use to understand the regulatory environment may themselves be unreliable. The story also highlights the growing operational challenge of distinguishing human-generated content from AI-generated content in high-volume digital environments.

What to review

Organisations that run digital consultation, feedback or engagement processes should review what controls exist to detect automated, coordinated or AI-generated submissions. This might include rate limiting, CAPTCHA or verification steps, anomaly detection on submission patterns, or qualitative review of responses for signs of templated or machine-generated language. For organisations that participate in regulatory consultations or rely on consultation outcomes to inform business decisions, it is worth considering how AI-generated manipulation of those processes might affect the reliability of the signals they produce. Building awareness of this risk within public affairs, communications and research functions is a practical starting point.

Source: The Guardian