Michael C. Bromby

To be Taken at Face Value? Computerised Identification

Scientific evidence such as fingerprints and blood, hair and DNA samples are often presented during legal proceedings. Without such evidence, a description provided by the victim or any eyewitnesses is often the only means to identify a suspect. With the advent of closed circuit television (CCTV), many crimes are now recorded by cameras in the public or private domain, leading to a new form of forensic identification--facial biometrics. Decisions on how to view and interpret biometric evidence are important for both prosecution and defence, not least for the judge and jury who must decide the case. A jury may accept eyewitnesses as reliable sources of evidence more readily than complicated forensic or scientific evidence. False eyewitness accounts appear reliable when confidently presented to a mock jury. The decision-making process of the judge and jury may be seriously flawed if an eyewitness has made a genuine mistake. Using computerised recognition, the judicial decision of whether to accept an alibi or whether to accept the eyewitness account may be helped by removing the inherent uncertainty of human recognition.

An Examination of Criminal Jury Directions in Relation to Eyewitness Identification in Commonwealth Jurisdictions

This paper presents an analytical review of judicial directions to guard against wrongful convictions based upon erroneous eyewitness identification evidence. Factors known as the Turnbull Rules, derived from the English case R v Turnbull,1 are of significance within many common law jurisdictions when considering the accuracy of eyewitness identifications and the practice of jury directions or mandatory warnings. The influence of these rules, together with variations in the approach taken by Commonwealth jurisdictions, illustrates that while the factors identified in Turnbull are to be found in the approaches adopted across the various jurisdictions studied, there is diversity in terms of whether or not such directions are mandatory and also as to their form and scope. The frailties of eyewitness evidence are of primary concern to any reliable prosecution and exist irrespective of jurisdiction. Such evidence has been the subject of psychological evaluation and the findings of such studies cross national, ethnic and jurisdictional boundaries. These findings, and the processes employed to test such evidence, can be used to develop the form which directions to a jury may take. The requirement for a jury direction and the scope of such a warning are two matters where there are differences between the approaches taken by individual Commonwealth legal systems. While approaches differ according to jurisdiction, generally either legislation and/or case law will provide guidance to judges in terms of formulating a sufficiently robust warning. It is suggested that this wide variation in practice is incompatible with the universal theory of honest, yet mistaken, eyewitnesses. However, any approach which is adopted must also seek to prevent the balance tipping so far in favour of the accused that a jury discounts the evidence of the honest and accurate eyewitness. This paper will examine the approach to eyewitness identification evidence adopted in several Commonwealth nations and evaluate the influence of the English Turnbull case within these jurisdictions. A universal paradigm will be suggested, which would require directions to be given to juries in all cases which involve eyewitness evidence. Such a direction would incorporate a common framework which the judge would then tailor as required to take account of the circumstances of each particular case.

A CommonKADS Representation for a Knowledge-Based System to Evaluate Eyewitness Identification

The Joseph Bell Centre for Forensic Statistics and Legal Reasoning has been set up to examine the correct presentation, interpretation and evaluation of scientific and forensic evidence through the use of technology. The aim of the Centre is to build computer systems for those operating in the legal system so that they can follow best practice whether investigating a crime or presenting evidence in court. The initial approach to developing computational systems is to build small-scale knowledge-based systems in specific domains. This paper presents a CommonKADS approach to designing a small-scale system to evaluate eyewitness evidence. CommonKADS is a Knowledge Acquisition Design System using computer-generated models to represent how tasks are performed, which agents are involved, their expertise and the communication involved in the process of evaluating eyewitness evidence. The knowledge to be modelled for the application has been drawn from sources such as: the police, the prosecution service, lawyers and psychologists. This system will be piloted and evaluated by the Centre with collaborating institutions for ultimate use by law enforcement agencies and prosecution and defence agents.

Identification, trust and privacy: How biometrics can aid certification of digital signatures

Public key infrastructure (PKI) enables the secure and private exchange of data using an unsecure public network, such as the Internet. The use of paired private and public keys, issued by a trusted third-party authority, enables documents to be transferred securely and for the sender to be authenticated. The use of biometrics offers the potential to enhance considerably the PKI model in restricting the use of your private key for encryption and decryption. The use of a fingerprint, for example, can provide a higher level of confidence than the traditional password/PIN model. This provides the additional level of individual or personal authentification should a group of people have access to one key. The authentification of data, or a document, is often physically remote from the owner, especially for Internet-based communications. Conversely, traditional biometric usage has been to identity the physical presence of a person, for example for secure entry, or the receipt of information, or the receipt of goods. Within the EU, the European Electronic Signature Standardisation Initiative (EESSI) has led to a plethora of standards covering PKI, electronic signature algorithms, electronic signature formats, time stamping, the provision of certification services, information security and the preservation of evidence. This paper illustrates how a legally compliant and secure framework for the verification and non-repudiation of digital technology can be established using PKI and biometric technologies. In particular, the legal requirements for digital signatures and their certification must be defined, especially with reference to biometric methods for certificate protection and access.

Security against crime: Technologies for detecting and preventing crime

The papers in this issue are linked under the broad theme of crime, of which several arose from the 20th Annual BILETA Conference in Belfast, April 2005 Technology has, of course, presented new approaches to both preventing and detecting crime—likewise, the counterpart ability to commit crimes is extending domestically and internationally. It is estimated that crime costs the UK tax payer £50 billion per year, with investment in technology consuming a significant portion of the budget. Academic research into crime prevention and detection is increasing and supported by the government via funding council projects such as the EPSRC Projects Against Crime.

Over-Observed? What is the Quality of CCTV in this New Digital Legal World?

Abstract Poor quality CCTV evidence requires a witness to make an identification of the suspect. Good quality CCTV images are often left to the jury, who are unfamiliar with the face. Psychological research has demonstrated that identification procedures for both known and previously unknown suspects should be very different. A visual comparison of high quality imagery may suggest that a more reliable decision regarding identity will be made. However, studies indicate that when an assailant is unknown, identification is poor even when the image is of high quality, regardless of format. As such, research has demonstrated that recognizing or matching unfamiliar faces even in optimal conditions is an extremely error prone process. This paper examines the current legal framework for identification from imagery in the light of psychological research. Incorrect identifications may not necessarily be safeguarded against in some situations, although further research is needed to elucidate reliable identification methods for criminal prosecutions.

Brain imaging: consent issues, and licensing of storage and analysis.

Notes for the Information Technology Think Tank Meeting, University of Edinburgh, 11th December 2009The UKs Human Tissue Act 2004 requires holders of human tissue to be licensed and subject to regulation by the Human Tissue Authority. The counterpart legislation in Scotland (Human Tissue (Scotland) Act 2006) does not enact a similar licensing regime north of the border and holders of human tissue are subject to far less onerous or strict regulations.Brain imaging does not, however, fit within the broad definition of human tissue as the scan (whether this is numerical data or a pictorial representation of the data) contains no actual tissue. However, analogies can be drawn with the concept of DNA testing as it is frequently the DNA profile (i.e. numerical data or diagram illustrating the data) that is stored and subsequently analysed, although the sample from which the profile is determined may also be stored. Brain imaging cannot be undertaken from a ‘sample’ so here the analogy between brain imaging and DNA cannot be made.