UK Home Office's Use of AI in Asylum Processes Under Scrutiny

The AI tools used by the UK Home Office to process asylum applications have come under the scrutiny of the Independent Chief Inspector of Borders and Immigration due to concerns over accuracy and transparency. A report prepared by the Open Rights Group highlights a lack of oversight in the systems used by the ministry, while the institution's refusal to share information regarding these tools has also prompted action from data regulators.

The 'Asylum Case Summarisation' (ACS) and 'Asylum Policy Search' (APS) systems under review utilize OpenAI's GPT-4 model and Microsoft Foundry infrastructure. These tools, which began pilot implementation at the end of 2024, were designed to summarize interview transcripts and provide case officers with information on country policies. The Open Rights Group argues that the efficiency gains provided by the systems overshadow the transparency regarding how errors are identified.

The Home Office rejects the allegations, stating that AI does not make final decisions and that the system only provides support to trained case officers. Meanwhile, the UK's data regulatory body is conducting a separate investigation after the ministry refused to disclose information regarding staff training and the use of the tools. Although the Inspectorate has not yet launched a formal investigation, it has requested additional information from the digital rights organization, taking into account concerns regarding the reliability of the systems.