UK borders watchdog weighs probe into Home Office GPT-4 asylum tools

The potential investigation comes after a six-month campaign by the Open Rights Group, which used 11 parliamentary questions, two ministerial letters and 10 public records requests to pry details from a department that has withheld them. Its findings challenge the official pilot evaluation: the share of ACS summaries deemed unfit for use—often for factual errors or misrepresentations—is significantly higher than the published 9 percent error rate, and there is no quality control for APS summaries.

The tools were introduced in late 2024 as the Labour government sought to clear a backlog of around 100,000 asylum cases. The department has since halved the number of people waiting on an initial decision in the year through June, but appeals against decisions have ballooned. The government's own AI Playbook requires departments to list algorithmic tools on a public repository; these two are not on it, and the Public Law Project says it has found 11 high-risk AI systems across government that are likewise absent.

The Information Commission is already investigating the Home Office's refusal to tell ORG how staff use and were trained on the tools, and lawyers commissioned by ORG argue that not being able to contest an incorrect AI summary may breach UK data rights laws. OpenAI, for its part, says it is working to improve model accuracy and will retire GPT-4 in January.