Fast Start for World's Biggest Biometrics ID Project

Identity is easy to take for granted. Most of us have multiple legal documents or ID cards that prove we are who we say we are. But for many of the world's poorest citizens, the lack of legal identity is a barrier to participating in commerce and receiving services.

In India, an estimated 500 million people have no form of reliable identification. It's a problem the Indian government has set out to fix through a five-year project with a budget of US $430 million for this year. Starting six months ago with rural populations, the government has begun to create a biometric database that will eventually contain an unprecedented hundreds of millions of records. "We are talking about 10 times more than anything else that has been done before," says Anil Jain, an IEEE Fellow and distinguished professor at Michigan State University, who is an expert in biometrics.

From each volunteer participant, the government collects 10 fingerprints, 2 iris images, and a photo, and if the new data don't match any identity already enrolled, it assigns the person a unique 12-digit number. After that, a single fingerprint or iris scan should be all that's needed to verify the identity of any person. As of the end of March, the Unique Identification Authority of India (UIDAI) has registered more than 4 million people this way. The UIDAI hopes to eventually collect biometrics from a majority of the Indian population.

India has many federal and state programs to help people living in poverty, but today it's nearly impossible to be sure that funds and benefits are actually being delivered to those who need them. The ID project is an attempt to cut down on fraud and graft by increasing accountability and transparency. It's also meant to provide access to banking and the formal economy that many people lack.

Government biometrics programs have been tried before and failed, in India and elsewhere. The United Kingdom's universal ID program, for instance, got bogged down by both costs and privacy concerns and didn't offer tangible benefits to the average citizen. But the UIDAI's universal ID program, or Aadhaar, as it's called, seems to be off to a fast start. As soon as he was appointed in July 2009, chairman Nandan M. Nilekani set the ambitious goal of issuing the first million IDs within 12 to 18 months, and the UIDAI hit that mark by January 2011. Efficiency is not a strength of most government bureaucracies, so Nilekani looked to Silicon Valley for help.

A core group of Indian expats with Silicon Valley start-up experience began working on the problem, as unpaid volunteers. A three-bedroom flat in Bangalore served as the group's living and work space for six months. "We converted the living room into an office and started designing the core of the biometric system," says Salil Prabhakar, a biometrics expert who was recruited for the project. In the meantime, Nilekani's office prepared the paperwork and logistics to give the team official authority.

Prabhakar and his colleagues started by looking at best practices developed by other biometrics researchers. The U.S. National Institute of Standards and Technology, the International Standards Organization, and the FBI "have done a lot of work to come up with good specifications [PDF]," says Prabhakar. "If UID happened 10 years ago, life would have been much harder for us." By using standard specifications, the Indian project has been able to use many existing devices and data interchange formats and avoid having to rely on a single private biometrics equipment vendor or proprietary format.

Still, the Indian project has presented unique obstacles. Take fingerprints, the most mature and best-understood biomarker. Because almost all scholarly publications had focused on Western subjects for law-enforcement purposes, Prabhakar and his colleagues were in the dark about how well such systems work when data are collected in a variety of locations and conditions from a rural population of Indians, many of whose fingerprints have been obscured or erased by manual labor.

So UIDAI started with smaller proof-of-concept programs to test the accuracy of the ID generation process with tens of thousands of subjects. There are two factors that determine accuracy: the false positive rate, which is how often a newly registered person's record is judged to be a duplicate of someone else's record, and the false negative rate, which is the frequency that true duplicate records—for instance, if someone registered twice under two names—are not recognized as such. In the pilot study, the researchers found that by adding the iris scans to the 10 fingerprints, they could decrease the false negative rate by a factor of 50 over the use of irises alone and by a factor of 25 over the use of fingerprints alone. "It's hard to make a system error free," says Jain, "but the important thing is to minimize the error."

When a duplicate record is detected, it is flagged for manual verification. In the pilot study, the system generated false positive errors in 0.0025 percent of cases. That rate would generate 25 erroneous records daily, if the project makes the goal of 1 million IDs per day by October 2011 set by Finance Minister Pranab Mukherjee.

In order to scale up so quickly, UIDAI has piggybacked on existing government infrastructure: Biometrics data are collected at government offices by government employees or private enrollment agencies. These workers use officially approved sensors costing $1000 to $2000. A single sensor can collect data from 50 people per day.

So far, the project has done better with speed than accuracy. Raj Mashruwala, part of the core volunteer team, says that the biggest challenge has been consistently getting high-quality data from the enrollment operators. "I don't think they are yet at a stage where it's working flawlessly," he says, but UIDAI will be updating training and procedures to make improvements.

All of this is still just the first phase of the program. Additional pilot tests are ongoing to figure out the best way to run verifications against the database. Despite the rapid early pace, it will be years before the project can succeed in all its goals.