Yesterday was the 38th and last day of hearings in the Supreme Court case challenging the constitutional validity of India’s biometric authentication programme. After weeks of arguments from both sides, the Supreme Court has now reserved the matter for judgement.
Since its inception, the Aadhaar project has lurched from controversy to scandal. In the last two years, the debate has heavily centred around issues of data security, privacy and government overreach. This debate, unfortunately, like with most things Aadhaar, has been obfuscated in no small part due to the manner in which the Unique Identification Authority of India (UIDAI) reacts to critical public discussion.
As India waits for the apex court’s judgement, this is as good time as any to take stock of the security and privacy flaws underpinning the Aadhaar ecosystem.
Poor security standards
Let’s start with the lackadaisical attitude towards information security. As has become evident in the past, harvesting and collecting Aadhaar numbers – or acquiring scans and prints of valid Aadhaar cards – has become a trivial matter.
There are several government websites which implement Aadhaar authentication while at the same time lack in basic security practices such as the use of SSL to encrypt user traffic and/or the use of captchas to protect against brute-force or scraping attacks. This includes the biometric attendance website of the Director General of Foreign Trade, the website for the National Food Security Mission and the Medleapr website.
With numerous government websites being susceptible, problematic issues such as the use of open directories to store sensitive data gives us a look into how even the bare minimum – when it comes to adhering to security best practices – isn’t enforced across the gamut of websites which interface with Aadhaar.
It should not be acceptable practice to have government websites with open web directories containing PDF scans of dozens of Aadhaar cards available for just about anyone to view and/or download. Yet, over the past year and even before, many government websites have been found to either inadvertently or knowingly publish this information without much regard for the potential consequences it could have.
The UIDAI has repeatedly shown an attitude of hostility and dismissiveness when it comes to fixing security and privacy issues which are present in the Aadhaar ecosystem. It has also shown no signs of how it plans to tackle this problem.
A man goes through the process of eye scanning for the Unique Identification (UID) database system. Credit: Reuters/Saumya Khandelwal
In my personal experience as a security researcher, I have found and reported a cache of more than 40,000 scanned Aadhaar cards being available through an unsecured database managed by a private company, which relied on those scans for the purposes of verifying and maintaining records of their customers.
What’s worse is that the media reports regarding Aadhaar information being exposed may only be scratching the surface of the issue as more data may actually be susceptible to access and theft, and simply yet to be found and publicly reported. For example, data could be leaking through publicly available data stores of third-party companies interfacing with Aadhaar, or through inadequately secured API and sensitive portals without proper access controls.
Not all security incidents become a matter of public knowledge, so what we know at any given point about the illegal exposure of Aadhaar information may just be a glimpse of what is actually out there.
It should be acknowledged that the possession of these 12-digit numbers and their corresponding demographic information can open up room for potential fraud – or at the very least make it easier for criminals to carry out identity theft and SIM and banking fraud.
A detailed analysis of all publicly-reported Aadhaar-related or Aadhaar-enabled fraud over the last few years shows that the problem is not only real but deserves far more attention than what it has received so far.
Threat level infinity
Taking a step back, it’s clear that the Aadhaar project snowballed into an ecosystem that it now struggles to control.
For instance, demographic information – as is stated in the draft for the Aadhaar Act (NIDAI Bill 2010) – was originally considered confidential information, meaning no entity could request your demographic information such as name, address, phone number etc. for purposes of eKYC.
However, as the ecosystem has progressed, the implementation and usage of eKYC have also changed and grown significantly with companies like PayTM utilising Aadhaar-based eKYC for the purposes of requesting and verifying customer information. It should be considered that data which has been collected by any of these companies through Aadhaar can be accessed by them in the future for an indefinite period of time depending on their own policies regarding storage and retention of the data.
If there ever is a breach of the CIDR or a mirrored silo containing a significant amount of Aadhaar-related data, it would directly affect more than one billion people. To put this in perspective, it would easily be the single largest breach of data in terms of the sheer number of people affected and it would have far-reaching consequences for everyone affected which might be very hard to offset.
On a comparatively smaller scale – although just as serious, if not more in terms of potential implications – would be a breach of any given state’s resident data hub (SRDH) repository. In some cases, SRDHs have been known to integrate data acquired from other sources containing information regarding parameters such as caste, banking details, religion, employment status, salaries, and then linking the same to residents’ corresponding Aadhaar data.
Damage control would be costly and painstaking due to the number of people enrolled. What adds to the disastrous consequences is that one cannot just deactivate their Aadhaar or opt-out of the programme the way they would with say a compromised Facebook or Twitter account. You can always deactivate Facebook. You cannot deactivate your Aadhaar. It should be noted that even with biometrics set to ‘disabled’, Aadhaar verification transactions can be verified through OTP.
Additionally, the Aadhaar ecosystem is such that information about individuals can be accessed not just from UIDAI servers but also from other third-party databases where Aadhaar numbers are linked with their own respective datasets. Due to this aspect – multiple points of failure are introduced for possible compromise of data, especially because third-party databases are almost certainly not as secure as the CIDR.
Aadhaar is designed to be a number, not a card. In fact, the word ‘card’ does not figure even once in the Aadhaar Act. Credit: PTI
Recently, after taking a closer look at the ecosystem of websites which incorporate the use of Aadhaar based authentication, I discovered that it was possible to extract the phone number linked to any given Aadhaar through the use of websites which poorly implemented Aadhaar text-based (OTP) authentication.
This process worked by first retrieving the last four digits of the phone number linked to an Aadhaar using any website which reveals this information (this includes DigiLocker, NFSM.gov.in and seems to be standard practice which seems to be enforced by UIDAI) and then performing an enumeration attack on the first six digits using websites which allow the user to provide both their Aadhaar number and the verified phone number linked to it.
This again highlights that while secure practices might be followed by the UIDAI, the errors in implementation and other flaws are introduced neverthelessby third parties who interface with Aadhaar, posing a risk to the privacy and security of its data.
The bank mapper rabbit hole
As of February 24, 2017, it was possible to retrieve bank linking status information directly from UIDAI’s website without any prior verification.
However, after this information was reported, the ‘uidai.gov.in’ website was updated to first require requesters to prove their identity before retrieving Aadhaar bank-linking data from the endpoint on their website.
A year later – when business technology news site ZDNet published their report regarding a flawed API on the website of a state-owned utility company (later revealed to be Indane) – part of the data revealed included bank linking status information which was identical to what was previously revealed on UIDAI’s website without proper authentication.
This suggests that both the Indane API and UIDAI website utilised the National Payments Corporation of India (NPCI) to retrieve bank-linking data – but as of now, this remains conjecture since Indane never put out a statement or gave a public comment regarding the flawed API on their website.
More importantly, what this also suggests is that the NPCI never placed any controls or security mechanisms (such as request throttling or access controls) on the lookup requests it processed for the UIDAI (and seemingly for Indane as well).
This means that while the UIDAI may have fixed their website to not reveal bank linking data without proper verification – the issue was not rectified at its core by the NPCI – allowing the same to happen a year later in Indane’s case. This practice also classifies as a case of security through obscurity, which “is the belief that a system of any sort can be secure so long as nobody outside of its implementation group is allowed to find out anything about its internal mechanisms”.
Who is on the hook?
There is a lack of needed accountability when it comes to data breaches. Have any of the organisations against whom allegations of data breach been made been investigated and acted on? Have fines been imposed on those responsible for allowing access/theft of user data? Have there been reports published by any of the affected organisations in which they investigate any alleged breaches to either provide insight regarding the breach and its impact, the scale of data accessed, logs of access and other crucial evidence or dismiss the allegations by proving that there was no intrusion which took place?
Most of the times, organisations do not even accept that a breach has taken place, let alone take responsibility for the same and strive to better protect user data in the future.
Switching to ‘PR spin mode’ should never be the answer when dealing with the data of billion-plus Indian citizens and residents. This can be observed in almost all cases where a breach or security lapse was alleged.
The UIDAI has also acquired the dubious reputation of sending legal notices and slapping cases on journalists and security researchers who seek to highlight the security and privacy problems ailing the Aadhaar infrastructure.
In March 2017, a case against Sameer Kochhar – chairman of the Skoch Group – was filed on the basis of a complaint from Yashwant Kumar of the UIDAI allegedly for “spreading rumours on the internet about vulnerability of the Aadhaar system”. Kochhar had written an article in February 2017 titled “Is a Deep State at Work to Steal Digital India?” in which a request replay attack on biometric Aadhaar authentication was demonstrated.
Two months later, The Centre for Internet and Society published a report regarding several government websites which were inadvertently leaking millions of Aadhaar card numbers. A few days after this report was published, the UIDAI sent a legal notice to the organisation, stating that the people involved with the report had to be “brought to justice”.
In January 2018, an investigative story was published by Rachna Khaira of The Tribune newspaper – in which she reported that access to an Aadhaar portal was being sold by “agents” for as cheap as Rs 500. In response to this story – the UIDAI first sought to discredit the investigative work by calling it a ‘case of misreporting’ – after which they attempted to downplay the magnitude of the report by citing that biometrics were safe and had not been breached.
Following this, the Delhi crime branch registered an FIR against the reporter and others named in the article on the basis of a complaint by a UIDAI official, with charges ranging from forgery, cheating by impersonation and unauthorised access of a computer system.
In March 2018, ZDNet published a report about Aadhaar-related data leaking from an unsecured API on a utility provider’s website. This was the result of days of testing to first confirm the existence issue and its scope. It was preempted by more than a month of attempted communication through several channels of communication – email, phone, even direct messages via Twitter – with both Indane and the UIDAI (and even the Indian Consulate in New York).
But still, when the report was published after a lack of acknowledgement/response from affected parties, the UIDAI was quick to deny the report as well as any possibility of such a thing occurring. The Aadhaar agency then released a statement in which they said they were ‘contemplating legal action’ against the publication of their report.
Data security and privacy laws won’t do much to affect the dismissive and hostile attitude the UIDAI seems to have regarding the people that investigate and report on security and privacy issues relating to Aadhaar.
Hide and seek
In general, when it comes to reports of security breaches and security incidents, many authorities in India prefer playing the blame-game. This was seen latest in response to an internal letter (ironically marked as ‘SECRET’) that was circulated on social media – which mentioned that data was stolen from the Aadhaar Seeding portal of the EPFO by hackers exploiting a known vulnerability in the Apache Struts framework.
Following this – the EPFO quickly switched to PR mode and publicly issued a statement through their official Twitter account (@socialepfo) denying the breach – saying that “There is no leak from EPFO database. We have already shut down the alleged Aadhaar seeding site run by Common Service Centres on 22.03.2018.”
Every time reports of a potential breach or leak of data circulate, Indian government agencies are quick to come out and announce that no breach has taken place. However, this is always to be taken just on the basis of their saying so, as opposed to the reports which they’re meant to be arguing (in some cases) contain verifiable evidence which is the result of arduous investigative work.
Regardless, passing around the blame and in cases completely denying security incidents is not something authorities should be doing when it concerns the data of more than a billion people.
In response to a recent story by Asia Times regarding Aadhaar enrolment software being cracked and sold, the UIDAI sought to discredit and discount the report through messages shared on their social media profiles – where they stated that the report was “baseless, false, misleading and irresponsible”.
Every time reports of a potential breach or leak of data circulate, Indian government agencies are quick to come out and announce that no breach has taken place. Credit: Reuters
The UIDAI should have an interest in protecting any and all data which stems from or relates to Aadhaar as it has to do with a project they are ultimately responsible for. It should not matter whether the leak occurred from a portal on EPFO’s website, an API without proper access controls on Indane’s website, a website of the Andhra Pradesh state government, through biometric request replay attacks, through sold access to admin portals and cracked software, or however else. It should ultimately be the UIDAI’s responsibility to not only be reactive about these issues when they’re brought to light but to do so in such a way which does not hinder reporters from continuing their work.
Additionally, if the UIDAI wishes to keep its systems as secure as they could be – they should proactively seek such reports about flaws or vulnerabilities in critical infrastructure pertaining to their project.
The way forward
In April 2018, the head of the Indian Computer Emergency Response Team (CERT-IN), rather defensively noted that “not a single person had reported any incident” to the organisation.
CERT-In, a part of the IT ministry, is the central agency responsible for dealing with security issues and incidents. To put it bluntly, it has not done a very great job of outreach when it comes to the people it ultimately relies on: security researchers and hackers.
In India, there is an abundance of skills and talent when it comes to IT security and this could be of immense help to organisations responsible for managing critical infrastructure – but only if they cared enough to utilise it to the fullest extent.
Ajay Bhushan Pandey, the CEO of UIDAI, promised a secure and legal bug reporting environment for the Aadhaar ecosystem sometime in 2017. However, almost a year later, there are no tangible signs of any steps being taken to ensure the same. In fact, the UIDAI would already be straying from their usual course of action if they stopped harassing people reporting on issues of security and privacy with regard to Aadhaar.
It has been suggested that the UIDAI employ a bug bounty programme – which involves rewarding hackers with monetary compensation or through means such as an addition to a ‘Security Hall of Fame’ as an incentive.
I personally believe that there is no need for a bug bounty programme in its traditional sense – meaning that UIDAI should not have to provide material incentives to attract hackers to report valid issues to them. Simply acknowledging the work of those that discover and report valid issues should more than likely be incentive enough to get talent on-board.
The US Department of Defense (DoD) employs a similar approach where they invite hackers from the world over to test their systems for security vulnerabilities/bugs and then report them in a responsible manner. What the hackers get in return is the acknowledgement of their skill and devotion to ensuring the security of DoD’s platform. Something similar needs to be set up with regard to critical information infrastructures in India so that issues can be reported by anyone who wishes to do so – without hassle and/or fear of persecution hanging over the heads of hackers.
Karan Saini is a security analyst and consultant. He writes on issues relating to web security, privacy and open source intelligence on his blog (@squeal).