Traditional understanding of privacy may not suffice in an age of digital identities. Supporters and opponents of UIDAI are missing the point
Written by Subhashis Banerjee | Updated: July 5, 2017 12:05 am
In the Aadhaar debate, strident dogmatic positions have far outnumbered credible peer-reviewed analyses, and the proponents and opponents have mostly talked past each other. There have been alarming reports of exclusion and disruption in social welfare but it is still unclear whether they are due to fixable teething troubles, careless deployment or something more fundamental, and what proportion is affected. The opponents too have been unable to make precise how exactly Aadhaar violates privacy, nor have the UIDAI and ministerial proclamations declaring Aadhaar to be perfectly safe engendered confidence.
Moreover, trivial and easily fixable examples of privacy breaches have been turned into big issues. Whatever were the initial plans, the government clearly wants to use the unique identification of Aadhaar to enforce compliance in a variety of schemes by avoiding duplicates. The opponents want Aadhaar to be voluntary and if that makes it a lame duck instrument, so be it. And the potential benefits of Aadhaar beyond de-duplication, for example in analytics, have not even been discussed much.
The disagreement has inevitably shifted to the courts. However, in the recent Aadhaar-PAN linkage case, several arguments from both sides were specious and not well analysed. No expert was examined and the judgement — though perhaps fair under the circumstances — did not inspire confidence in the process.
For example, the petitioner’s argument on legislative competence — that the linkage cannot be made mandatory in the IT Act without first removing the contradiction from the original Aadhaar Act — appeared to be compelling. Yet the court dismissed it. However, even if the court had upheld it, the objection was more on procedural grounds and not fundamental in nature, and at best the government would have been forced to go back and amend the original act.
The petitioner’s arguments under Article 14, that the mandate discriminates between different classes of taxpayers, must have sounded tenuous even to the petitioners and were summarily rejected.
The argument that PAN cancellation violates the right to practice any profession was accepted, but so were the state’s arguments on the need for de-duplication. The court also accepted, without question or calling for any analysis, the state’s assertion that biometric de-duplication is perfect. Partial relief was given to non-Aadhaar holders on the ground that cancellation of PAN will cause hardship.
The petitioners had put forth another set of problematic arguments based on dignity and bodily autonomy, on the state’s right of eminent domain over the human body and on informational self-determination. The court deferred them for consideration by a larger bench, along with all issues related to privacy. The mandated use of Aadhaar for IT is egalitarian, and any perceived indignity of fingerprinting is due to prejudice. Moreover, fingerprints and iris scans (both can be contact-less) are fundamentally no different from facial photographs; they are images and not parts of one’s body. They can be used for matching and de-duplication either manually or automatically. They differ only in efficacy and not in principle. Unfortunately, the response from the state — claiming that the state indeed has a right over the human body — was irrelevant and disproportionate.
The question then is: Can the state insist on an identification mechanism? If so, under what circumstances? What are the limits of informational self-determination? Note that the state has already assumed this right, many years back without much protest, by making PAN cards with photographs mandatory for tax returns. The purpose even then was de-duplication, only the methods and their efficacy were different.
So, the main issue is privacy, which the court has been deferring, and little has been said on it to enable an informed decision. On the one hand, the state’s position that Aadhaar is safe because UIDAI stores only minimal data required for biometric matching and demographic details, is untenable. The government and UIDAI cannot absolve themselves of the responsibility of protecting users from privacy breaches through possible correlation attacks on linked databases. Further, the possibilities of insider attacks also need to be considered.
On the other hand, the opponent’s claim — that collecting biometric information and storing them in a central database and linking multiple databases through the Aadhaar number fundamentally violates privacy — is also without any careful evaluation of a precise threat model. For example, PAN cards are already linked to bank accounts, ITR and major purchases. How does linking Aadhaar increase the possibilities for correlation attacks? Why is making the Aadhaar number public more dangerous than making PAN public? Biometric and demographic details are publicly available anyway, and anybody determined enough can obtain these from touched objects and using a powerful camera even without the victim’s cooperation.
Clearly, it will be unsafe to use biometrics for authentication, to access bank accounts for example, but what about only for identity verification and de-duplication? Surely we need to exhaustively enumerate the possible ways in which privacy may be compromised and model an attack surface? Only then can the questions related to privacy protection, either through technical or legal means, even be asked. The assertion that privacy protection is impossible with biometrics and a global ID is far from established.
It may not be enough to apply a traditional understanding of privacy to the new scenarios presented by digital identity and the internet. The need of the hour is for our institutions to wake up and carry out conservative, detailed and rigorous analysis of all issues involved — social, economic, technical and legal. Till then, it will be best to go slow with Aadhaar, engage, analyse, correct, and ensure that there are no hardships.
The writer is professor, department of computer science and engineering, IIT Delhi