UK Watchdog May Probe Home Office AI Use in Asylum Cases — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- ICIBI is considering an investigation into the Home Office’s use of AI tools for asylum processing, after requesting information from Open Rights Group about their deployment and reliability
- Home Office uses Asylum Case Summarisation and Asylum Policy Search tools built on OpenAI’s GPT-4 and Microsoft Foundry to summarize interview transcripts and country policy data, aiming to cut case review time by about an hour per application
- Open Rights Group found the AI summary error rate is significantly higher than the Home Office’s reported 9 percent, with no quality control in place for policy search outputs, raising concerns about factual inaccuracies in high-stakes decisions
- Labour MP Bell Ribeiro-Addy and Public Law Project warn that cultural bias in GPT models could distort asylum narratives and lead to discriminatory outcomes, especially when nuances in persecution accounts are lost
- Information Commissioner’s Office is already investigating the Home Office over its refusal to disclose how staff use and are trained on the AI tools, following public transparency mandates under the UK’s AI Playbook
- Public Law Project identified 11 high-risk AI systems across government not listed in the official public repository, suggesting widespread unreported use of automated decision-support tools
Why it matters: Asylum applicants face potential harm from unverified AI summaries that may contain errors or bias, while the Home Office’s non-disclosure violates transparency rules and risks breaching data rights laws—undermining public trust in a system already under legal scrutiny.
Ask SkimNews
