UK home office under scrutiny for AI use in asylum decisions

A UK government watchdog is considering an investigation into the Home Office’s use of generative AI tools for asylum decisions, citing concerns over accuracy and secrecy. The Independent Chief Inspector of Borders and Immigration (ICIBI) has requested more information from the Open Rights Group (ORG), which conducted a six-month review. The Home Office uses Asylum Case Summarisation (ACS) and Asylum Policy Search (APS), built on OpenAI’s GPT-4 model, to summarize interviews and country policy. A Home Office spokesperson said AI tools assist, not replace, human decision-makers. Researchers and campaigners, including Labour MP Bell Ribeiro-Addy, raised concerns over AI accuracy, especially in high-stakes cases. ORG found unfit summaries were higher than the published 9% error rate. The Home Office has not disclosed tool use to asylum seekers, raising data rights concerns. The AI Playbook mandates transparency, but tools are not listed publicly. Caroline Selman of the Public Law Project noted other high-risk AI systems are also used without public disclosure. OpenAI plans to retire GPT-4 in January, while the Information Commissioner investigates the Home Office’s refusal to disclose staff use of the tools. Campaigners stress the need for transparency and accountability in AI-driven asylum decisions.