Implementing a national digital ID biometrics-based programme - Hindustan Times
close_game
close_game

Implementing a national digital ID biometrics-based programme

ByHindustan Times
Aug 12, 2022 02:45 PM IST

The article has been authored by Alok Tiwari - VP & BU Head - Public Security & Identity, IDEMIA India

Before the pandemic, there was a general initiative from governments around the globe to implement remote governmental services so that citizens could renew ID documents, fill in their tax returns and complete other basic tasks without in-person interaction with the administration. However, the Covid-19 pandemic has revealed that governments globally are far from their aim of becoming fully digital. It has also been demonstrated that preventing online identity fraud will be one of their main challenges in the coming years.

The pandemic has shown that remote governmental services such as applying for and renewing ID documents, filling out tax returns, requesting repeat medical prescriptions, accessing online telemedicine appointments and voting are necessary for the functioning of society.(Vipin Kumar/Hindustan Times)
The pandemic has shown that remote governmental services such as applying for and renewing ID documents, filling out tax returns, requesting repeat medical prescriptions, accessing online telemedicine appointments and voting are necessary for the functioning of society.(Vipin Kumar/Hindustan Times)

Now let us consider the fact that we all have a secure way to prove our identity: Our biometrics. Easy, accessible and secure, biometric face verification offers a reliable and convenient method of checking that citizens are who they claim to be. Before governments embark on this journey, we have identified three key success factors that must be considered.

Hindustan Times - your fastest source for breaking news! Read now.

Remote governmental services―a necessity for the functioning of States

The pandemic has shown that remote governmental services such as applying for and renewing ID documents, filling out tax returns, requesting repeat medical prescriptions, accessing online telemedicine appointments and voting are necessary for the functioning of society. It is vital that governments provide their citizens with e-services, and the only way to do it is through a digital ID system that is secure and easy to use. The pre-requisite for citizens to access e-services is to be able to easily and securely prove their identity, and this is where biometrics step in.

The aim is clear: Offer remote services through a national digital ID scheme facilitated by secure face verification. But what are the key ingredients to make such a complex project a success? How can citizens be reassured that their data is safe? And how can citizens and the private sector be encouraged to embrace the project for its ultimate success―the widespread adoption and the true digitisation of private and public sector services?

Three key success factors that governments must consider before they embark on a digital ID programme.

One, the securest authentication and the highest protection of citizen personal data. The widespread adoption of any national digital ID scheme will only happen if all users involved are convinced that their personal data is secure and that the authentication technology is accurate. We have already clarified that biometric technologies offer an accurate and convenient method of authenticating the identity of a person. To ensure that an individual is who they claim to be, using one form of biometrics is a good idea, but combining two or more reinforces accuracy. Multibiometrics, for example, verifying a person’s face along with their fingerprints, prove beyond a shadow of a doubt that they are who they claim to be.

We need to adopt attack detection methods to prevent fraudulent access to governmental services. Fraud is a serious issue; therefore, biometric algorithms need to continually be strengthened and improved in order to stay one step ahead of fraudsters to thwart all attempts of ID fraud.

Attack detection techniques offer additional reassurance for all parties involved in a national digital ID system. Face verification technologies often include a passive and active approach to prevent spoofing attacks. The passive approach consists of taking a selfie to prove that they are actually presenting their face to the camera (and not a picture or a mask). The active approach relies on interaction between the user and the application, aka the ‘challenge request’. Several methods are available on the market such as nodding, blinking and smiling. Both the active and passive approaches ensure that the person is alive and is not an impostor wearing a mask.

Efforts must be made to ensure personal data security for the user. Nearly all users will access remote governmental services using their smart device (smartphones and tablets). However, there are certain questions being raised regarding the security of accessing government portals and sharing sensitive data using a personal smart device. From a citizen’s point of view, their smartphone has many benefits; it is convenient, always accessible and personal data stored on their smartphone remains under their control. Despite this, there is some criticism to the reliability and security of a smartphone. For example, it is common for users to download third-party apps that are not always thoroughly vetted and may render information on a smartphone vulnerable. However, secure data sharing and processing techniques make it possible to offset these disadvantages.

Depending on the approach chosen for the national digital ID scheme, the user’s personal and biometric data can remain on their smartphone, which is where the verification techniques are carried. The smartphone completes a verifiable calculation that it shares with the organisation that requested it without ever divulging any of the user’s personal data. This technology, called verifiable computing, means that one central entity can outsource the computing of data to another potentially unknown, not previously verified entity, while maintaining verifiable results. This means that citizens can do the matching of their own data to verify their identity on their smartphone (i.e., the unknown, not verified entity), without anyone doubting the validity of the computing done. Citizens control their biometric data at all times, and it never leaves their device.

Multi-party computation provides another layer of security. Whether the verification is built into the app or is centralised in the cloud, with multi-party computation, the processing of the data is shared by different parties. All the open, vulnerable data is not processed by one central player but several contributors. Only by breaching the data processed by each player would the data make sense to a malicious perpetrator. The right technology provider will offer both technologies for the national digital ID scheme.

Two, the right technology provider should be used. Finding a reliable technology provider to protect governmental services with biometric technologies can be a daunting task. Checking that the selected provider has worked with governments and private sector service providers worldwide and has a proven track record is comparatively easy. However, how do you know which provider can be trusted with citizens’ personal data? Furthermore, algorithms can be tricky to understand. How do you know if the algorithms are working as they should, and if they are doing exactly what the technology provider states?

The simple answer is to always select a provider whose algorithms have been certified by independent third-party testers. To test the performance of algorithms, there are many independent bodies that provide neutral evaluations that are available to the public. It is very important to check how biometric technologies from different companies measure up against each other based on large data volumes.

In order to have the best algorithms on the market it is important for technology providers to continuously invest in improving them. One of the most important aspects of Artificial Intelligence-based automated face verification is to teach algorithms to be accurate, but also to be fast and optimised for fairness.

Citizens of a population differ in age, gender, ethnicities and facial features (beards, glasses, and even makeup). These differences produce a variation in algorithm performance, which means the more diverse the data used to train it, the better the algorithm will perform.

Biases primarily come from training databases, which is why technology providers need make sure that their databases contain a variety of images of the same element in various acquisition conditions while respecting local and international privacy regulations—for instance, European providers are subject to strict regulations when using citizen personal data.

Ensuring face verification algorithms are fair and unbiased toward everyone is not an easy process, but the most advanced deep learning solutions are now up to the task.

Three, ensure the best user experience. Another key success factor to encourage the widespread adoption of the national digital ID scheme by all parties, and thereby guaranteeing its success, is the ease of use of the technology involved. To ensure a digital ID is user-friendly, the face-capture environment has to be taken into account.

Face verification as a means of biometric identification to unlock phones, for example, has been in existence for many years, and, therefore, users are more comfortable with its usage. Yet, factors such as the lighting, time of day and even angle of capture cannot be predicted, which is why all remote governmental services need to include software that guides the user and provides easy-to-understand instructions.

The article has been authored by Alok Tiwari - VP & BU Head - Public Security & Identity, IDEMIA India

SHARE THIS ARTICLE ON
Share this article
SHARE
Story Saved
Live Score
OPEN APP
Saved Articles
Following
My Reads
Sign out
New Delhi 0C
Tuesday, March 05, 2024
Start 14 Days Free Trial Subscribe Now
Follow Us On