Open Rights Group accuses UK Home Office of embedding AI in secret, sham-marriage detection and phone scanning feed a digital hostile environment

A report by the Open Rights Group claims artificial intelligence has been embedded across the UK Home Office in near-total secrecy, deployed to detect sham marriages, estimate the ages of asylum seekers, scan the phones of new arrivals, summarise asylum interviews, and surveil borders with facial recognition, drones and satellites. The report, titled 'Automated Discrimination: Unpacking the Digital Hostile Environment', was assembled from the limited information the department publishes, freedom of information requests and questions from MPs.

The report describes the Identify and Prioritise Immigration Cases (Ipic) tool as opaque, and argues that AI used in immigration is neither neutral nor accurate but replicates existing bias. Its framing borrows deliberately: the "hostile environment" was Theresa May's phrase as home secretary in 2012, and the report claims successive governments have reduced complex human cases to a technical problem that can be fixed. "Relying on AI to decide who gets to stay here and who gets thrown back into danger is a deeply alarming move," it states. Officials routinely invoke national security to stonewall FOI requests and redact disclosures, the report adds.

It has drawn cross-party backing — Labour's Bell Ribeiro-Addy, the SNP's Pete Wishart and the Green party's Siân Berry. Ribeiro-Addy noted that the Home Office has been reluctant to publish documents explaining how AI is used in asylum processing, how problems are addressed, or why the most vulnerable people with the least power should bear the greatest risks. The Open Rights Group wants a ban on undemocratic AI uses — it specifically names ChatGPT tools in asylum applications — a public register of government AI tools, and a narrower definition of national security.

A Home Office spokesperson said AI is "just one of the tools at our disposal to protect our borders" and does not replace or automate human decision-makers, with asylum decisions made by trained caseworkers.

The same week, HMICFRS found that UK police and fire services failed cybersecurity inspections, with legacy IT rated the biggest threat and backup systems unsatisfactory. The department building an automated hostile environment for immigrants is the one whose own email cannot survive an audit.