Aadhaar that doesn’t exclude
- 11 April, 2018
University of British Columbia
Aadhaar is in the news today partly because of security concerns and partly because of reports that the poor are unable to receive PDS rations because of failures in Aadhaar authentication. In this article, Kotwal and Ramaswami focus on the latter with an eye to look for ways to bring down the exclusion errors.
At the outset, we should make clear that we were enthusiastic cheerleaders for Aadhaar until quite recently. Technology, we had hoped, would solve the problems of corruption as well as exclusion.
What has changed? Why the rethink?
We believed that like any new technology, Aadhaar would be experimented with, first in urban areas where there was good connectivity and with a clientele that was somewhat used to electronic transactions, and once the flaws had been troubleshot, then it would be tried in other areas. We never imagined that a state like Jharkhand, with a large tribal population and weak connectivity, would have Aadhaar Assisted Biometric Authentication (AABA) imposed on it before the reliability of the technology had been established. Inevitably, there have been multiple reports of authentication failures affecting the most vulnerable segments of the population in the most vulnerable areas. Ironically, these are the people who have the greatest need for subsidised food. Technology thus deployed defeats the very purpose of the Food Security Act. The starvation deaths reported in these areas highlight the seriousness of the problem but they do not reveal its magnitude in terms of the number of people adversely affected.
A recent study by Jean Drèze and his co-authors, based on a survey of about 1,000 households in 32 villages in Jharkhand, estimates that the exclusion errors were as high as 20% in areas where biometric authentication was required for every sale. This is simply unacceptable.
The response of the UIDAI (Unique Identification Authority of India) to such reports is blanket denial. This would carry some conviction if the denials were accompanied by official data on exclusion errors. The UIDAI, of course, maintains that it only provides a platform and that applications (such as PDS) that use Aadhaar are within the domain of the respective ministries. Even if true, it does not justify the absence of active efforts by the UIDAI in locating the extent of exclusion errors and the contribution of authentication failures.
In recent days, there has been a flurry of announcements on the use of virtual IDs to secure greater privacy and the use of facial biometrics to reduce exclusion errors. The willingness to improve the technology is laudable. But for the immediate, the UIDAI must confer with the line ministries and the state governments to reflect on the design of the PDS. The limits of technology, whether in terms of the failure to authenticate biometrics or the absence of connectivity, should not hurt the poor.
The contradiction of the current design is that it places the PDS dealer – from whose corrupt dealings the poor were to be saved by Aadhaar – firmly in charge of making biometrics work for each beneficiary every month. This does not have to be so. Some years ago, the Madhya Pradesh (MP) government devised a scheme where the biometric authentication would be done only once a year (and that too by agencies unconnected with the PDS). On that basis, households were to be given coupons to be cashed at fair-price shops. The coupons were bar-coded to prevent duplication and misuse. The commissions to administer the scheme were indexed to volume. So administrators had every reason to minimise the exclusion error. The beauty of this scheme was that it placed manageable demands on Aadhaar authentication and did not rely on the PDS dealer to make it work.
While the MP government lost its nerve on the eve of its implementation, the government and the UIDAI must pick up on good ideas and be realistic about how and when Aadhaar authentication can be insisted on. They should realise that while pursuing the problem of identity fraud, they are creating a far more serious problem of exclusion errors in areas where the human costs of exclusion are the highest. The government should seriously consider ‘Aadhaar light’ designs such as the MP scheme.
We don’t see why there is the need for biometric authentication for every sale when the information about the account holders (Aadhaar numbers) as well as transactions are computerised. If the point-of-sale terminal has a digital record of all the customers (along with their Aadhaar numbers) entitled to buy their rations in that outlet, there is little scope for identity fraud. A fake card would show up as a duplicate. Why not just issue Aadhaar holders smart cards that can be swiped like credit cards rather than biometric authentication which requires a far more sophisticated technology that we have not been able to operate reliably yet? As long as the payment to the dealer is contingent on the recorded transaction via a smart card, like any other credit card transaction, the scope for identity fraud is minimised.
Of course, there remains the possibility of quantity fraud whether the authentication is done through smart cards or biometrics. The PDS dealer can just lie to an ignorant customer and assert that the authentication had failed and
A technology is only as good as the judiciousness with which it is used.
“Aadhaar that doesn’t exclude” by Prof. Ashok Kotwal and Prof. Bharat Ramaswami. Reprinted from THE INDIAN EXPRESS with the permission of The Indian Express (P) Limited © 2018.
- Nazar Khalid
- 10 January, 2018
- Sohini Paul
- 24 July, 2016