UK Facial Recognition Immigration Rules: How It Works and Fails?

The Home Office has published updated guidance governing Immigration Enforcement's use of live facial recognition (LFR) to locate people of immigration law enforcement interest, setting out rules for watchlists, deployment locations, human oversight and biometric matching thresholds. The policy, dated October 2026, extends beyond ports of entry to transport hubs, travel routes and events, and provides for joint deployments with police forces across the United Kingdom. Daily Dazzling Dawn understands the document instructs Immigration Enforcement (IE) staff on "how to use overt LFR to locate relevant people lawfully and ethically" while supporting "IE's and the Home Office's legitimate immigration enforcement aims."

The guidance traces a step-by-step chain from lawful purpose to accountable deployment. Defining lawful purpose comes first; only then is it decided who is eligible to appear on an LFR watchlist. Watchlists must be linked to a defined deployment objective, location and context; time limited; and proportionate, with a necessity and proportionality assessment completed before any operation proceeds. Use must be overt, and human officers retain the decision on engagement or enforcement — the technology flags a possible match, but people act on it.

Images of people who do not generate an alert are "automatically and permanently deleted," and deployments must be "monitored, reviewed and accountable," according to the policy. IE's Emerging Technology Team, within Strategic Services and Transformation, provides senior internal oversight. Transparency is described as "the default position," with any departure from advance notice exceptional, justified and recorded — and not affecting the requirement that deployments remain overt and subject to post-deployment publication and scrutiny. Higher-risk deployments, including those at transport hubs and events, require additional oversight. Cooperation with police is also provided for: IE can supply watchlists for police facial recognition systems or process police-supplied data through its own systems, extending the technology's operational reach.

The guidance is candid that false alerts are inherent to the technology. Operational experience suggests the false alert rate (FAR) should generally be 0.1 per cent or lower — fewer than one false alert per 1,000 comparisons. The policy cites 2021 testing by South Wales Police and the Metropolitan Police Service to "test the accuracy and equitability of facial recognition tech in operational settings," and a report on the NEC system used by National Police Chiefs' Council forces. IE will continue to use a threshold of 0.64 for IE-led deployments using approved and objectively tested technology. The study also noted that "the system was more likely to recognise watchlist subjects as age increased" — a demographic bias finding that speaks directly to equitability concerns.

Daily Dazzling Dawn understands the residual risks sit in the seams of the system. A 0.1 per cent FAR still means real people are stopped in error when deployments process thousands of faces; the human decision point is the primary check, but it depends on officer judgement under time pressure. Watchlists are time-limited and objective-linked on paper, yet proportionality is assessed by the deploying authority itself. The transparency default permits exceptional departures from advance notice that are merely "justified and recorded," leaving room for operations the public learns about only afterwards. Joint deployments with police blur which force's safeguards govern when watchlists and data cross systems. And the age-related accuracy finding leaves unresolved questions about equal treatment that the threshold alone does not answer.

The document frames adherence as a trust mechanism: IE staff must follow agreed LFR policies and processes, which "helps IE use LFR lawfully and effectively, with appropriate safeguards, and helps maintain trust and confidence among the public, partners and other stakeholders." Whether that confidence holds will depend on how rigorously the review, recording and publication requirements are enforced in practice — and on whether the acknowledged error rates and bias findings are monitored as closely as the deployments themselves.