What procedure should be followed when police officers identify suspects?

Guidance on this issue has been provided by two recent decisions of the Court of Appeal.

In R v Smith and others (CLW/08/31/2), the court considered identification from CCTV footage. After a murder committed with firearms outside a nightclub by a group of men, officers were asked to view CCTV footage to see if they recognised any of the offenders, and identifications of various suspects were made by several officers. At trial and on appeal, several of these identifications were challenged.

Although dismissing the appeal for other reasons, the court observed that, although a police officer asked to view a CCTV recording in order to see whether he recognises anyone in it is not in the same shoes as a witness asked to identify someone he has seen committing a crime, the safeguards put in place by PACE Code D are equally important to such a case.

In particular, where an officer does recognise someone in such a situation, there must be in place some record which assists in gauging the reliability of that assertion of recognition, and so it is important that the officer’s initial reactions to the recording are set out and available for scrutiny. These should include whether he fails to recognise anyone on first viewing but succeeds in doing so subsequently, what words he uses by way of recognition, whether he recognises anyone else, whether he expresses any doubt as to the recognition, and what it is about the image that he says has triggered the recognition.

The court said that it is therefore vital that a protocol should be prepared to provide a safeguard for measuring the recognition against an objective standard of assessment, otherwise there can be no assurance that an officer is not merely asserting that which he wishes and hopes, however subconsciously, to achieve, viz, the recognition of a guilty participant.

In R v Flynn and St John (CLW/08/33/5), the court considered voice recognition evidence. After an armed robbery during which the offenders used a van to enter premises, officers identified the two defendants from audio recordings made covertly inside the van before the robbery. Again, these identifications were challenged at trial and on appeal.

The court said generally that voice recognition identification is more controversial than visual identification and more likely to be reliable when carried out by experts using acoustic and spectrographic techniques and sophisticated auditory techniques rather than by ‘lay listener identification’, which requires that the witness possesses some special knowledge of the person whose voice he purports to identify.

The ability of a lay listener to identify a voice is subject to a number of variables, including:
    (a)the quality of the recording of the disputed voice.
    (b)the gap in time between the listener hearing the known voice and his attempt to recognise the disputed voice.
    (c)the ability of the individual to identify voices in general.
    (d)the nature and duration of the speech which is sought to be identified.
    (e)the familiarity of the listener with the known voice.

It said that even in optimum conditions, a confident recognition by a lay listener of a familiar voice may be wrong, and that the key to admissibility with lay listener evidence is, therefore, the degree of familiarity of the witness with the suspect’s voice, which will depend on the facts of each case.

The court went on to say that, where lay listener evidence is given by police officers, it must be treated with great caution. An expert should be instructed to give an independent opinion on its validity and a careful record should be kept of the procedures adopted, which should comply with certain minimum standards.

It also held that all voice recognition evidence (expert or lay) must be accompanied by a careful direction to the jury warning of the danger of mistakes, and that where the ident