AI analysis of police body-camera footage surfaces underreported stops and racial disparities, raising constitutional and evidence integrity concerns

A new study published on 7 May 2026 analysed thousands of officer-worn camera recordings using AI and found evidence of underreported police stops, troubling racial disparities in officer interactions, and constitutional concerns about how AI-based analysis of body-camera footage is being used in law enforcement. The research identified patterns in officer behaviour that were not captured in official reports, raising questions about the reliability of human-generated documentation compared to AI-assisted video analysis. The findings add to growing evidence that AI video review tools can surface discrepancies between official police records and what body cameras actually recorded, with direct implications for evidence integrity, officer accountability and the admissibility of AI-generated analysis in criminal proceedings. For EU law enforcement, the study is relevant to AI Act Annex III requirements for transparency and human oversight of high-risk AI systems used in law enforcement decision support, and to the ongoing debate about whether AI-assisted body camera analysis constitutes a biometric or behavioural surveillance system requiring conformity assessment.