The rapid rise in accessibility and portability of cameras has resulted in widespread reliance on the interpretation of images by analysts and investigators in criminal proceedings. Codes of practice, guidance and jurisprudence have evolved to facilitate the admission of opinions as to the identity of offenders (or persons of interest) at trial. In this article we explain why allowing investigators to give opinions as to identity on the basis of familiarity with images or suspects acquired during the course of an investigation is incompatible with mainstream scientific research and advice, and conducive to error. It rests on the flawed assumption that investigators can reliably identify or recognise persons in images, articulate and document the basis of these “identifications”, and avoid the risk of contamination (really cognitive bias) from their knowledge of, or exposure to, domain-irrelevant information. Jurors, who may be invited to conduct their own comparison between an image and the defendant in the dock, are similarly vulnerable to assuming the task is straightforward, as well as contextual and cognitive biases. Using the facts and evidence in R v Yaryare  EWCA Crim 1314 as a case study, we show how case information available to investigators and imaging analysts both informs their interpretations of images and is (re-)presented at trial and on appeal as independent support for their opinions. We identify substantial threats to fairness, proof and rationality and propose that only witnesses with demonstrable expertise should be permitted to testify as to the identity of persons of interest in images.
|Journal||The Journal of Criminal Law|
|Publication status||Accepted/In press - 4 Feb 2023|