APPENDIX 1
Memorandum from the Government
1. SCIENTIFIC
ADVICE, RISK
AND EVIDENCE
IN THE
IDENTITY CARDS
PROGRAMME
This section covers the mechanisms by which
conclusions are arrived at and the external sources of evidence
and guidance used by the Identity Cards team. It will also explain
how these data and the conclusions drawn from them are presented
to the programme's assurance groups and to ministers.
Risks relating to biometrics
1.1 Biometrics is a key technology underpinning
the delivery of the programme's aims (although it should be noted
that the scheme's integrity is not wholly dependent on biometrics
because of the biographical or "footprint" checks which
will be made on applicants) and so examples on the use of scientific
evidence in the Identity Cards programme will be drawn from this
area.
1.2 Some biometrics technology is relatively
immature and the whole field is fast-moving. The programme anticipates
a biometric scheme comparable to the few, large-scale deployments
which currently exist there is a necessity to base our models
of the likely performance of the biometric technology on the best
scientific evidence. Further, because it is a key dependency,
it is necessary to have the best possible assurance of the work
done on biometrics and also to ensure that the risks and the potential
of biometrics are properly communicated to ministers.
1.3 The uncertainties over biometric technology
and thus those areas which need most attention are encapsulated
in several key risks associated with biometrics. These are summarised
below:
— It may be impossible to prevent applicants
falsifying (spoofing) their biometrics. This risk can be mitigated
through analysing the threat posed and designing the correct detection
processes and by ensuring that the deterrent regime is appropriate.
— The matching of newly enrolled biometrics
against all those already enrolled may not be 100% reliable, raising
the risk that a very small number of people may be able to enrol
more than once without authorisation[1].
Likewise, the matching of an individual's fingerprint or iris
against their previously enrolled biometrics in order to verify
their identity may not always be reliable. Mitigation of these
risks requires modelling the likely distribution of, and the mechanisms
behind, these mismatches and then looking at the impact on stakeholders
and modelling the ways in which these mismatches will be detected
by other, non-biometric, means.
1.4 Within the context of these risks, key
requirements of the biometrics need to be set:
— Which biometrics should the scheme
use, and will these be sufficient to meet the performance demands
of the scheme?
— What are the criteria and tests we
should apply to biometric technology suppliers?
1.5 It is not necessary for the purposes
of the programme to embark on publicly-funded scientific research
to improve the capabilities of biometrics. Our approach to answering
these questions has focused on gathering the best evidence available
from existing programmes, academia and suppliers, and on using
trials to answer specific questions for which data are not available.
Sources of evidence, advice and assurance
1.6 We undertook a comprehensive survey
of the academic and commercial literature on biometrics, and the
published reports available from existing biometric schemes, government
laboratories and standards bodies.
1.7 In 2003 NPL undertook a feasibility
study for us on the use of biometrics in a national identity card
scheme. One of the recommendations of this report was the trial
we ran in 2004 to gather evidence on public perceptions and attitudes
towards biometrics and data on the timings of biometric recording
and verification processes. This was conducted with UKPS and demonstrated
that the overall experience was positive and met or exceeded the
expectations of a vast majority of people. It provided some findings
on the technology, but was not an assessment of the technological
capabilities of biometrics.
1.8 We are currently funding technical work
at NPL to define a methodology to assess the interoperability
of fingerprint systems using systems from the four leading fingerprint
identifications systems. We are planning a set of trials for 2006
which will look at the relative performance of suppliers' integrated
biometric systems in the areas of timing, usability and spoofing-resistance,
and the relative performance of suppliers' matching algorithms
on a "test" database of biometric records.
1.9 We are working with leading biometric
experts from:
— US National Institute for Standards
& Technology.
— UK National Physical Laboratory.
— Members of the Biometrics Working
Group.
— Communications-Electronic Security
Group (part of GCHQ) and other experts in the field of fraudulent
use of biometrics.
1.10 We are learning from the experiences
of our colleagues from:
— UK Passport Service (UKPS).
— Immigration & Nationality Directorate
(IND).
— Police Information Technology Organisation
(PITO).
— International Biometric Identity
Cards Schemes such as Hong Kong & the Philippines.
1.11 We have consulted with leading biometric
technology firms in industry:
— The Identity Card Programme has met
with 15 companies involved with biometrics and is currently undertaking
further market sounding activity in relation to biometric matching
performance. This contact has taken place in accordance with OGC
(Office of Government Commerce) procurement rules.
1.12 We receive assurance on biometrics
from a number of sources, principally:
— We have given evidence on biometrics
on two occasions to the Home Office Science and Technology Reference
Group (chaired by the Permanent Secretary).
1.13 The government's Biometrics Assurance
Group and the Home Office Senior Biometric Advisor will review
biometric aspects of the Identity Cards Programme.
— Sir David King, the Government's
Chief Scientific Adviser, chairs the Biometrics Assurance Group
which has been established as a panel of internationally eminent
specialists in biometrics and related technologies.
1.14 Scientific evidence and its meaning
for the programme and the output from the programme's assurance
bodies are presented to ministers via submissions and face-to-face
briefings. Numerical data is frequently presented (eg the current
measured levels of passport fraud as a percentage of applications
and the anticipated reduction as a result of the introduction
of biometrics) as are comparisons with other schemes (both those
in other countries such as the US, the UAE, the Netherlands, Hong
Kong, the Philippines and Germany and also relevant schemes in
the UK such as IDENT1, the Police fingerprint database).
January 2006
1 There will be exceptional circumstances (eg for the
purposes of Witness Protection) under which people will be permitted
to enrol more than one identity. Back
| 
