The UIDAI has so far detected 34,015 cases where one person has been issued two Aadhaar numbers
Surabhi Agarwal
First Published: Tue, Mar 05 2013. 11 40 PM IST
Against an enrolment of 290 million people till date, the duplicates represent a little more than 0.01%.
New Delhi: As it moves closer to covering one-fourth of the country’s 1.2 billion people under the unique identity, or Aadhaar, project, the Unique Identification Authority of India (UIDAI) has detected 34,015 cases where one person has been issued two Aadhaar numbers.
UIDAI officials said the percentage of duplicates is well within the authority’s projected estimates calculated through proof-of-concept studies and accuracy tests in the past.
Still, the revelation does raise some concerns given the centrality of the Aadhaar number to several key programmes.
Against an enrolment of 290 million people till date, the duplicates represent a little more than 0.01%. To be sure, this is lower than the 0.035% of duplicates estimated by an accuracy study by UIDAI after 84 million enrolments.
Even at accuracy levels as high as 99.99%, there are likely to be a small number of duplicates, a UIDAI spokesperson said in an email to Mint.
“On current national population of 1.21 billion, the number of likely residents with two Aadhaars would be 1.21 lakh @ 0.01%,” the email added.
From a pure numbers point of view, the error percentage is very small and almost negligible, said writer and columnist Dilip D’Souza, who writes a column for Mint on math.
“Any project of this size is going to have some errors and even this tiny percentage of error in a country of our size will translate into thousands of people,” he said, adding that it would be unreasonable to expect 100% accuracy in such a project.
UIDAI added that based on the experience so far, “substantive number of residents tend to enrol more than once, contributing to more than expected number of duplicates”.
The other reasons were misuse of biometric exceptions (a provision for residents whose biometrics such as fingerprints and eyes don’t register definitively) and software application bugs, aside from the statistical probability of 0.01% of duplicate enrolments.
The authority said that while it is continuously working on upgrading systems and tightening procedures to address the first two causes, the last one can be “controlled by reducing the number of duplicate enrolments, which is currently around 5%”. This implies that a large proportion of such enrolments are detected by the system.
Both UIDAI and the census department under the National Population Register project are recording biometric data. Even though both the agencies reached a truce after a cabinet decision in January 2012 and were allowed to co-exist, there have been several reports of duplication between the two agencies in biometric collection.
UIDAI is not just being used as the main platform for rolling out the government’s ambitious direct cash transfer scheme, but is also becoming an important authentication factor for several security measures and financial transactions.
While the Aadhaar number has been notified as proof of address and identity for opening a bank account, it is also a valid know-your-customer norm for getting new telecom connections. Some states such as Delhi have also made Aadhaar mandatory for availing any government service along with marriage and property registrations.
Moreover, the basis of the United Progressive Alliance’s key electoral plank for the forthcoming general election in 2014—the direct cash transfers project—is the fact that its integration with the unique identity number will check payment leakages by weeding out fakes and duplicates from the system.
Nikhil Dey, co-convener of the National Campaign for People’s Right to Information, said he’s not surprised by the duplicate Aadhaar numbers.
“There is nothing like 99.99% accuracy and it just proves technology is not the magic wand which will solve everything,” he said.
Referring to direct cash transfers, Dey said that technology is being “pushed down the throat” and “it is a very bad idea as something like this needs 100% efficacy to be successful otherwise the system will collapse requiring manual overrides, etc.”.
Dey added that in India, be it the unique identity or any other project, achieving 100% accuracy is difficult and Aadhaar is using technology that’s globally untested on this scale.
UIDAI said in the email, “Thanks to the process of suo motu and voluntary check for duplicates deployed by UIDAI, we have been able to ensure high degree of accuracy while simultaneously not denying any resident volunteering to enrol himself/ herself.”
UIDAI added in its note that about 15 million duplicates had been detected by its back-end systems before Aadhaar generation and another 2.5 million possible duplicate records, for which the IDs have not yet been issued, are awaiting manual adjudication, along with the 34,015 mentioned above.
“In all these cases, the duplicate Aadhaar number has been cancelled,” it said.
One of the key reason for the duplicates slipping through is the collection of biometrics.
According to the accuracy study conducted after 84 million enrolments, around 0.23% of the residents’ fingerprints and irises didn’t register well, making the record vulnerable to error.
UIDAI added: “As the same study has brought out, a few thousand duplicates would still slip through the system by the time 121 crore resident were enrolled.”
The authority said it continuously reviews its applications and processes, besides monitoring accuracy standards and plans to incentivize the biometric service providers (BSP) to raise accuracy levels. “Going forward, it has been decided to motivate the BSPs to achieve high standards of accuracy by measuring the same and incentivizing them,” it said.
