Asylum: Artificial Intelligence
The question
To ask the Secretary of State for the Home Department, what steps her Department is taking to monitor outcomes from the Asylum Case Summarisation system to ensure that summaries are not provided to caseworkers if they (a) include missing information and (b) are inaccurate.
Answered by Alex Norris
Asylum Case Summarisation (ACS) was designed with the ‘Human in the Loop’ principle in mind. This means it is not possible for decision makers to use the tool to decide an asylum claim; it is an aid to case working. The tool uses a Large Language Model to extract and summarise information from existing asylum interview transcript documents to provide decision-makers with a concise summary document. Asylum Decision Makers are required to read all evidence and case notes specific to the claimant, before deciding a claim.
Part of the quality assurance framework asks if all documents submitted in support of a claim have been considered, with quality assurance checks regularly carried out on both asylum interviews and decisions.
As part of ongoing evaluation of the Asylum Policy Search (APS) tool, Subject Matter Expert (SME) testing continues in conjunction with the Country Policy and Information Team. A dedicated inbox was created for feedback and/or errors to be flagged and rectified. This is in addition to the existing quality assurance checks conducted in Asylum Operations. ACS has (as of today 27/04) become operational, and the same approach to ongoing evaluation and feedback mechanisms is being set out.
Verbatim has judged this answer against the question that was actually asked — answered, partly answered, or evaded. Sign in to see the verdict →