APS
2026 APS Annual Convention · 2026
Can Current Safeguards Protect Against Misinformation In an AI-Generated Police-Incident Report?
- Emily Spearing
The University of Western Australia - Levana Garreffa
The University of Western Australia - Kimberley Wade
University of Warwick - Deryn Strange
John Jay College of Criminal Justice, CUNY - Ullrich Ecker
University of Western Australia
Abstract
Generative AI is increasingly used to produce police-incident reports but may introduce errors. We found that such errors can distort incident memory and perceptions. Two safeguards were shown to provide partial protection against misleading AI content: actively reviewing the report to identify errors and watching the associated body-worn camera footage.
