A new biometric Face Verification Service (FVS) was launched on 16th November 2016. It will provide the Department of Foreign Affairs and Trade (DFAT) and Australian Federal Police (AFP) access to citizenship images held by the Department of Immigration and Border Protection (DIBP).
As per the fact sheet: “The FVS is a one-to-one image based verification service that can match a person’s photo against an image on one of their government records, such as a passport photo, to verify their identity. Often these transactions will occur with the individual’s consent.For example, where a person uses their citizenship record as evidence of their identity to apply for a passport, the system could enable the passport office to ask DIBP to confirm the identity of the passport applicant.”
The media release on the Australian Attorney-General Department’s (AGD) website talks about plans to expand the FVS service in the future to other government agencies, and to include images such as visa, passport and driver licence photos.
The FVS is expected to complement the Document Verification Service (DVS) in order to help combat identity crime.
There is a mention of technical ability to use still images from other sources such as CCTV, surveillance photography, the internet or social media, with the added caveat that law enforcement agencies can access those only with permission to do for a specific purpose under existing laws.
The next step is a planned launch of a Face Identification Service (FIS) in 2017 to help determine the identity of unknown persons. The FIS would only be used for investigation of serious offences, such as paedophilia or armed robbery. Access to it will be provided to a limited number of users in specialist areas. To help guard against the potential for false matches, they will be trained in interpreting the results.
The Australian Government has invested around $18.5 million in the development of these two Face Matching Services, citing that 4-5% of Australians (750,000 to 900,000) experience identity crime each year that results in financial loss of around $2.2 billion per year and a further $390 m for preventing and responding to identity crime (Source: Identity crime and misuse in Australia).
Several measures have been incorporated in order to allay security and privacy concerns over providing access to such sensitive personal information:
- Access criteria
The media release mentioned that the FVS’s privacy safeguards are based on independent privacy impact assessments (PIAs) conducted throughout the design and implementation phases.
The Face Verification Service Access Policy lists down a series of criteria for agencies to comply with, in order to gain access to the FVS services. The criteria encompasses conducting internal PIAs for information flow through FVS, entering into Interagency Data Sharing Arrangements (IDSAs), detail the scope of data sharing, creating a system for management and training of nominated users.
The AGD media release provides a template for the IDSA. For the initial launch, five FVS IDSAs are in place between DIBP and DFAT, and DIBP and the AFP.
- Not a centralised database
The system has adopted a ‘hub and spoke’ model, with the hub acting as a router to share images from existing databases between agencies so that they can undertake matching on a ‘query and response’ basis. The agencies match using their own facial recognition system and return results back via the hub. The hub does not conduct any matching and does not store any personal information. It only retains the de-identified transaction data that is necessary for monitoring the system.
- Audits and independent oversight
Participating agencies will continue to be subject o audits and independent oversight by existing external bodies, such as privacy commissioners, ombudsmen and anti-corruption or
- Security accreditations
Appropriate security accreditation, in accordance with the requirements of the Australian Government Protective Security Policy Framework and Information Security Manual from the Australian Signals Directorate, is required for IT systems, which are Nominated Users or connected to the Hub.
Agencies must provide Accreditation certificates to each other. Security accreditation must be informed by a Security Risk Management Plan (SRMP) which is regularly updated. The SRMP should be undertaken in accordance with the Protective Security Policy Framework, and a copy of the Plan must be provided to AGD.
Media release from AGD: Face Verification Service