86% Fake, 100% Admissible? Kennedys Law and S-RM Examine Deepfake Detection Evidence Standards for Courts

A joint analysis by Kennedys Law and digital forensics firm S-RM examines how AI-detection tools that output probability scores (such as ‘86% likelihood of deepfake’) create legal admissibility challenges because courts require explainable, repeatable methodology. The authors argue that forensic reports combining audio-visual cues with metadata analysis offer the most defensible approach.

In this domain: Sussex Police switch on Acusensus AI cameras for seatbelt and phone offences · Germany and France expand police AI powers without coherent regulatory concept – Verfassungsblog analysis · Metropolitan Police confirms Palantir AI pilot to flag officer misconduct patterns