Biometric Voter Registration (BVR)
Kenya has moved a step forward in boosting confidence in its elections systems with the introduction of the Biometric Voter Registration (BVR.BVR uses fingerprints and facial features to uniquely identify each voter. It incorporates total automatic data capturing with advanced technology known as OMR-ICR from captured digital images of voter forms.
The patented, accurate, personal and demographic data capturing technology affords a lowest costs solution with the auditable integrity of voter registration forms that are signed by the voters. It encompasses other data such as signature, fingerprints and personal picture. The complete and accurate voters list creates a stable foundation for a credible election.
BVR Eliminates Double Registration
A poll that is both credible must also prevent voters from voting more than once and unregistered voters from voting. This, the system does, through use of biometrics for automatic de-duplication, verification and authentication at the point-of-counting.
There are two main types of bio-identification systems, the visual bio-identification and electronic bio-identification.
Visual bio-identification methods include the use of photographs, signatures and fingerprints. Computers and/or registration staff then compare the images to detect possible double entries. The system is relatively cheap to implement and administer. On election day, polling staff compare the photographs and signatures on the voters’ identity cards with those on the voters’ roll. The human eye is still an excellent recognition aperture and a voter ‘s card with a reasonable picture provides a high level of security.
Technological gadgets can be used to capture handwritten signatures or fingerprints electronically. These digitised images are then sent to the Electronic Management Bodies (EMB) central database over a computer network where EMB staffs perform visual comparisons.
Digitised data can also be automatically analysed and compared using appropriate software. The software compares patterns in the digitised images and can flag possible mismatches for a human operator to investigate.
Biometrics is the technology by which the physical characteristics of a person’s face, fingerprint and iris are attached to the individual’s personal data and stored in the database. Future referencing of a person is based on this data. The system searches the existing database either to make a negative match or a positive match to a person’s stored biological data.
BVR by all means is a highly advanced biometric information system allowing one to enroll and identify millions of voters quickly and unmistakably. This is expected to be a smooth process as Kenya registers its more than 22 million voters.
Using the biometric identifiers, the possibility of election fraud is minimised, at the same time considerably accelerating the voter identification process. This has been a key worry bearing in mind that the last elections threw the country into a mess in 2007-2008.
Some of the most important features of the system includes a full range of biometric parameters which identifies the voters’ by fingerprints, iris and voice, highly customisable software modules for both input and output settings, comprehensive data logs and reports for comprehensive voter activity monitoring and the high-level security settings for data protection.
Fingerprint scanners are used in the BVR because they provide a fast, easy, efficient and secure measure through which, an individual with the proper access privileges can authenticate. The fingerprint of a voter is stored in a database that the scanner queries every time it is used.
There are two fundamental conditions the scanner goes through when an individual’s print is scanned. First, the print is usually searched for in a database of fingerprints. Once it is located, it then looks at the print to see what access privileges are associated with the print and compares them to the access they are trying to gain. If everything checks out, the subject is allowed access.
Another critical feature of the BVR kit is the finger printing recognition. This technology records and recognises fingerprints within seconds. In this case, the system is integrated with a microcontroller and other peripherals to form an embedded technique, which is a comprehensive electronic voting machine with fingerprint print identification system.
The OMR-ICR based solution that incorporates biometrics for voter authentication and duplication involves the following process:
• Preparation of the voter registration form data collection and management process
• Call for potential voters to register to vote
• Voter fills in form and signed affidavit
• Registration officers assist-proof voter to ensure completeness and accuracy
• The form is then scanned for entry of primary data. This will keep image of the original voter registration as part of the voter registration records
• Collection of biometric digital data which involves facial picture, finger prints and signature in the form of biometric or digital
• Printing of temporary paper card for the voter to verify correctness of the information posted.
• Constitution of registration database involving all voters registration data
• Verify and duplication. Check address against GIS, check fingerprint for duplicate registration and check others information and or databases for duplicate registration
• After that, the official voter registration cards are issued. Voters are mailed and notified of the card availability. They are then allowed to sign and pick the cards.
• Finally, the registration database is updates and a poll list, for specific locations, constituted.
Advantages of a BVR system of voter registration
BVR makes use of a unique form identifier. Each voter cataloging is embedded with both GIS and voting sub-divisional information along with unique identifier to prevent any human and machine errors for capturing the voter personal and demographic data. This prevents tampering or unauthorised changes for ultimate integrity.
The system will also save voter registration costs and error. The new technology reduces the need for clerical staff that manually types voter information when the voter personal and demographic data are captured automatically with the used of the patented and proven OMR technologies.
Another advantage is that all data including voter’s signature, fingerprints and pictures can be scanned and captured at the same time in forming part of the audit log and database of voter registration. This eliminates employee overtime costs and data entry concerns.
Fears about the BVR system:
Despite the promising outlook of the new voter technology, there are pertinent issues, which bring with it some weaknesses.
Network, systems and infrastructure security
There is no doubt that any network, system or infrastructure is prone to security gaps. In Kenya, as was the case with Ghana, there were no requirements for all systems to be configured according to security best practice requirements and audited to ensure compliance with the states best practices.
Since multiple critical files will be heavily used in this process, it is critical that File Integrity Monitoring will have to be deployed to detect changes to critical system and data files.
Physical security gaps
The BVR has not make it conspicuous of the controls to address how backups will be securely transported from polling centers to data centers. This means there is no mitigating control to prevent anyone from replacing the flash key backups during transport. In the incidence of data corruption of mobile registration toolkit the USB backups will be the only available source of data and if that is stolen, destroyed or replaced during transport then voter biometric template data is lost forever.
There are existing gaps when it comes to addressing cases where voter fingerprints cannot be captured, for example, people with missing fingers from injuries, diseases, etc. The IEBC has not provided any alternatives for verifying voters in the incidence that at that they have lost their fingers at the time of voting.
What options are in place to address a situation where the centralised voter registration databases crashes and cannot be recovered? How many levels of redundancies have been built into architecture to ensure that even if someone broke into data centre? Will backups be stored in multiple data centres to reduce such risks?
Failure to dictate age of voter
The process of BVR in many cases cannot detect minors and non-Kenyans and can work off of a system with missing or incorrect citizen data. The Ghana case where the BVR stations were not interconnected meant that there was risk of duplication. A case this March was pointed out where a voter registered 15 times. In total, more than 8,000 cases of multiple registrations were observed in Ghana early in the year.
For the last three years, the voting system has become popular in Africa with countries such as Nigeria and the Democratic Republic of the Congo initiating it. Recently, Sierra Leone also registered 2.5 million voters through the system. Work is in progress in Zambia, Namibia, and Mozambique to have BVR kits in place.
Is it worth the risk?
Depending on the infrastructure of a country and future developments in data transfer speed and availability, linking registration devices with the central database could provide high-level security as a new applicant can be checked instantaneously against the existing voter database.
Applicants not already recorded in the database could then be registered, thereby avoiding double registration. The system could also be connected online to the national population database, in which case applicants could only register to vote if they are positively identified in the national identity database.
For identification on Election Day, voters could be identified with biometrics before receiving their ballots. The voter could also be automatically marked in a real-time, live database as having voted and thus prevented from receiving another ballot anywhere else and attempting to vote more than once.
These scenarios are technically possible in Kenya, and with some further developments in computer speed and programme security are probably even feasible. There are, however, still technical and infrastructural constraints such as limited bandwidth, communication problems and lack of reliable power sources to using such systems in in many developing countries.
Given the complexity associated with establishing and running an electronic biometric voters’ register it is difficult to identify the exact benefits achieved and to justify the financial costs. Kenya will need to consider carefully whether an electronic biometric register really improves the level of security compared to a system where the voter signs for receipt of a ballot and has his/her fingers marked with indelible ink.