Policy brief explores safe use of speech recognition in policing

Posted on Monday 16 March 2026

Members of the Forensic Speech Services (FoSS) group in the Department of Language and Linguistic Science, in collaboration with the UKRI AI Centre for Doctoral Training in Safe AI Systems (SAINTS), have produced a new policy brief on the use of speech recognition in policing.

Published on 9 March 2026, the brief represents a consolidation of research and knowledge exchange across forensic speech science and the safe use of AI. It addresses the growing interest in AI technologies within policing, noting that some forces are already using speech recognition tools despite concerns over fairness, transparency, accuracy, and consequences in high-risk contexts.

The authors emphasise that the integration of these technologies needs to be carefully managed and rigorously tested, with human oversight remaining at the forefront of procedures. The brief outlines several key recommendations for policymakers to ensure safe and ethical deployment.

The brief, titled "Speech Recognition in Policing: Exploring the safe use of AI-based speech and language tools for law enforcement", is a joint effort. It was co-authored by FoSS researchers Jessica Wormald, James Tompkinson, Lauren Harrington, Ben Gibb-Reid, Philip Harrison, and Vincent Hughes, alongside SAINTS researchers Ibrahim Habli, Colin Paterson, Faisal Alkhatib, and Madeleine Nielsen.

Read the policy brief: https://drive.google.com/file/d/1s9_DAXm4brEnV9smW6LQk4C9fde12GB4/view