Centre for Internet & Society

The bill is basically the same as the UPA version, with some cosmetic changes, and some tokenism towards the right to privacy, says Abraham.

Shreeja Sen interviewed Sunil Abraham. The article was published in Livemint on March 8, 2016.

The government’s bid to push financial inclusiveness and access to government services has received a fresh boost, with finance minister Arun Jaitley introducing a proposed law to give legislative backing to Aadhaar, being implemented by the Unique Identification Authority of India (UIDAI).

This project, which uses a person’s biometric data like fingerprints and iris scans to authenticate identity of people receiving subsidies and other state benefits, will move India towards a cashless economy and help digital initiatives such as biometric attendance, Pradhan Mantri Jan Dhan Yojana, digital certificates, pension payments and the proposed introduction of payments banks.

Sunil Abraham, 42

Abraham is executive director of Centre for Internet and Society, a Bengaluru-based think tank focusing on accessibility, access to knowledge, telecom and Internet governance. He has written extensively on the UID scheme, and the intersection of privacy and security. He founded Mahiti—an enterprise that aims to reduce the cost and complexity of information and communications technology for the voluntary sector by using free software.

The Aadhaar project has faced its share of roadblocks with cases challenging it pending before the Supreme Court. A constitution bench of the court will decide whether the right to privacy is a fundamental right and if Aadhaar violates it.

Sunil Abraham, the executive director of Centre for Internet and Society, a Bengaluru-based policy research institute, is a critic of Aadhaar for several reasons. He explained his concerns in an interview. Edited excerpts:

Have any of the concerns regarding the Aadhaar project since its inception in 2009 been addressed?

Whatever we complained about six or seven years ago, whatever complaints were made by the civil society...all of those complaints remain in the exact same situation.

Nothing has changed.

What kind of concerns?

The first thing to remember is that privacy and security are just two sides of the same coin. You cannot have one without the other.

Our first concern with the project is centralization. Whenever you build an information system, and you create a central point of failure, then it will fail because the possibility of failure exists. The Internet has no central point of failure. That is why it is so difficult for you to bring the Internet down. Complaint number 2 is the opaque technology.

UIDAI keeps saying that “we have built a technology using a free software and open standard stack”. The first is a de-duplication software and the second one is the authentication software—those are the most important pieces of software.

This software is proprietary and nobody knows how they work and nobody can independently audit them.

The third complaint is the use of an irrevocable and non-consensual authentication factor. In the UID scheme, the biometrics serve two purposes: it can be used to identify a citizen and it can be used to authenticate a transaction. Authentication factors, commonly known as passwords, should always be revocable. That means if the password is compromised, you should be able to change the password or at least say that this password is no longer valid. The use of biometrics eliminates those two important requirements.

Further, in most other authentication, the process of authentication ensures that you are consenting. For example, PIN (personal identity number) authentications. But suppose I am authenticating you through your irises, then as long as your eyes are open, the machine will think you’re authenticating. There’s no way of saying I don’t want to authenticate. Or if you’re sleeping, somebody can hold your fingers over a biometric reader and open your iPhone. So that’s complaint number three.

The fourth complaint from the privacy perspective is: there is a very important database that they don’t talk about. I call it the transactions database. Suppose there is somebody who is using the UIDAI service to authenticate a transaction, then UIDAI should keep a record of that successful or unsuccessful transaction authentication. That means you have been registered into the database.

You go to a fair price shop to purchase subsidized grain and at that fair price shop or ration shop, you use your finger on the biometric reader, and then the UIDAI system says “yes you are indeed who you say you are”.

So, at that point, later the shop should not be able to say X never came here, or X came twice. So, in order for them to not say all those things, a record should be made on the UID database, that on this day, from this geographical location, this particular biometric reader sent us X’s biometric template and asked if the template matched against X’s UID number...the transaction database can be used for profiling. They never talk about it.

They never tell us what that database holds and how long they’re keeping all those records. None of that is clear.

Does Aadhaar bill help assuage your doubts about the project?

The government narrative has not changed in the last six years; the bill is basically the same as the UPA (United Progressive Alliance) version, with some cosmetic changes, and some tokenism towards the right to privacy. The proof that the technology is fallible is in the bill.

If the technology was infallible, as the UIDAI would like us to believe, then the bill would not criminalize the following: (1) impersonation at the time of enrolment; (2) unauthorized access to the Central Identities Data Repository.

Imagine that the bill admits that every Indian’s biometric can be stolen from one single centralized database. Now why don’t we have a similar offence for stealing all private keys from the Internet—we don’t because that is technical impossibility thanks to decentralization.

Therefore we don’t need a law to make (it) illegal. We’ve suggested changes to both the technology and the law. We’ve written seven open letters to the UIDAI, and we’ve never gotten any response. Very few of our concerns have been addressed. We’ve seen dogs getting UID, various other things getting UID, so there’s a lot of evidence that the system does not work. From Kerala we have stories of one person getting several UIDs, so we have no idea about technological feasibility of the project.

One of our distinguished fellows, Hans Varghese Mathews, has published an academic paper in the latest EPW (Economic and Political Weekly), by extrapolating UIDAI field trial data to national scale. He predicts that by the time the number crosses 1 billion, every time UIDAI tries to register someone new, they will match with about 850 people already in the database positively. So, the unique identification capability of the UIDAI will not scale above the billion. The consequence of the technology failing is not trivial. If someone replaces your biometrics in the central database, then the onus is on you to prove that you are a resident of India.

Previously, human beings determined the answer to this question, and they had to find proof that you were not a resident. Now, a fallible technology will be asked to answer this important question.

Isn’t the basic function of the Aadhaar project to ensure that benefits reach the person they are meant for, and it’s easier for people to get an identity proof for those who have no other ID, like migrant workers?

Two responses: is it good anti- corruption technology? Unfortunately not, because it is intended at retail fraud. The person under surveillance is very poor. But the person responsible for corruption is not poor. So, I believe you should be surveilling those responsible for corruption.

What I had said is UID should be first given to every single bureaucrat and every single politician in the country. From Delhi till the Panchayat office, till the ration shop in the village, that supply chain must be monitored and documented using cryptography, so that nobody can deny anything. We need non-repudiatable audit trail from New Delhi to the village because according to all analyses, that is where the theft is happening—in the supply chain. The villager who is taking false benefits, that is called retail fraud.

The bulk of the fraud is actually wholesale fraud. Please tackle wholesale fraud using non-repudiatable public audit trail from New Delhi to the village first, before you start surveilling the poor.

The second point is that people find it easy to get the UID. That is fine, but there is a problem; that it’s not uniquely identifying anybody. So, people will keep registering and the UID system will keep giving them more and more UIDs because there are no human checks and balances. Because you’ve gone with a pure technological solution, it’s very easy to fool (the system).

So, the ease of registration has not served the purpose.