Open Rights Group alleges secrecy around Home Office immigration AI in new report
The group wants stronger disclosure rules and bans on rights-threatening uses. The Home Office says trained caseworkers—not AI—decide asylum claims.
Loading page…
The group wants stronger disclosure rules and bans on rights-threatening uses. The Home Office says trained caseworkers—not AI—decide asylum claims.
Listen to this story
Open Rights Group’s report identifies tools used or tested for asylum casework and border monitoring, and argues that national-security exemptions have left their operation and risks largely hidden from public scrutiny. The Home Office says its asylum systems assist trained caseworkers rather than deciding claims, and that it monitors performance. A separate government timetable says facial age estimation was not operational as of May 29, 2026: testing is planned through 2026, with possible border use in 2027.
The report says the Identify and Prioritise Immigration Cases tool, or Ipic, helps assess and prioritise asylum cases.
Other uses it identifies include summarising interviews and country-policy information, live facial recognition, and AI-powered drones and satellites.
The Home Office says facial age estimation would use a photograph to estimate age, not identify someone against a database, and would inform rather than replace officers’ judgment.
Open Rights Group alleges that AI is being embedded across UK immigration and border work with little public scrutiny. Its new report, covered by the Guardian, links tools for asylum casework and border surveillance to a wider accountability problem: officials invoking national security to restrict disclosures about systems used in sensitive immigration work.
The report, Automated Discrimination: Unpacking the Digital Hostile Environment, draws on limited Home Office disclosures, freedom-of-information requests and questions from MPs. It argues that AI trials are taking place in near-total secrecy, with national-security exemptions used to block requests or redact information.
Its concerns extend beyond whether a machine makes the final asylum decision. The group describes AI handling information, helping assess cases and supporting surveillance. It says the technology can reproduce social bias and discrimination, particularly when its deployment is driven by anti-immigration policies and rhetoric.
Labour MP Bell Ribeiro-Addy criticised the department for withholding documents explaining AI’s role in asylum processing, its problems and how those problems are being addressed. She also challenged it to explain accountability when tools fail, and why vulnerable people should bear the risks.
Asylum decisions are made by trained caseworkers, with AI tools used only to assist in navigating interview records and policy information.
Home Office spokesperson, quoted by the Guardian
The Home Office’s response rejects the idea that its asylum tools replace or automate human decision-makers. It says they cannot be used in isolation and do not decide claims. The department also says it regularly monitors their performance through its established quality-assurance processes.
For age assessment, the Home Office’s guidance published on May 29, 2026, distinguishes testing from operational use. It said facial age estimation was not then operational, with testing planned throughout 2026 and possible implementation at the border in 2027. That timetable qualifies the report’s broader description of AI involvement in determining asylum seekers’ ages.
The proposed technology estimates age from a facial photograph; it does not search a database to identify someone. The guidance presents it as supplementary information for officers, not a definitive age determination or a replacement for their judgment. It acknowledges that an incorrect initial decision can place a child in an adult environment, or an adult among children.
Open Rights Group wants more than assurances about human oversight. It is calling for AI regulation and bans on uses it considers undemocratic or threatening to human rights, including ChatGPT tools in asylum applications. Its proposed transparency measures include stronger freedom-of-information laws, a narrower definition of national security and a robust public register of government AI tools.
ORG programme manager Sara Alsherif urged strict regulation to restrict AI use and prevent wider rollout, arguing that migration is a human experience rather than a technical problem. The Home Office, meanwhile, says it has already committed to publishing updated evaluations of the tools’ rollout. Its stated next step is further evaluation disclosure, not the restrictions the group is seeking.
Loading discussion...
Join the conversation
Explain where human oversight reassures you—or falls short.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.