South Africa’s plans for a national digital identity system could represent a leap forward for national safety, security and public service delivery. However, the effectiveness of such a system will depend on its ability to determine the liveness of the faces presented for verification.
The move towards a digital identity system is a logical one − prompted largely by the ease with which criminals can commit identity fraud using traditional identity documents. Physical ID books and cards are prone to abuse because they are relatively easy to forge or amend, and because they are so widely used and shared.
They are scanned at virtually every access point to secure estates and office parks, copies are stored by service providers and retailers, and identity numbers are frequently targeted by hackers.
Clearly, our ID numbers are not secret anymore. They are stored with every single bank account, medical aid and insurance policy. Whether we like it or not, our ID numbers have probably been bought and sold many times over.
But what if that didn't matter? What if the only thing that is going to determine whether this transaction is successful or not is verification of a person’s live face at the point of a transaction?
The critical key to securing a national digital identity system is to ensure that the liveness of any ‘selfie’ can be trusted.
Given that there are constant penetration attacks on financial institutions and people are being conned on a daily basis to part with passwords and account numbers, the best way to secure a transaction is with a live picture, to verify one’s credentials against a national digital identity database.
Digital identity based on biometrics − specifically facial recognition − promises to simplify the identification of individuals, speed up processes and reduce the risk of identity theft. There are myriad use cases for digital identity based on facial recognition.
Once the verified digital identity has been onboarded, public service agencies, banks, retail, hospitality and education sectors could use the technology to improve public safety, reduce the risk of fraud and fast-track digital service delivery.
Advanced facial recognition tools can even be used to verify age, gender, or geographic origin, to track sentiment, or to monitor public spaces for known suspects.
However, even faces can be copied − especially in an era in which anyone’s likeness can be copied off social media channels, or AI can be used to create a ‘selfie’ or deepfake video.
We need to acknowledge that the criminals will always try to find ways to circumvent new security measures. In fact, AI-assisted impersonation and spoofing is already reported to be soaring worldwide, with a Datos Insights report saying AI threatens to significantly amplify the threat of synthetic identity fraud to bypass biometric verification systems. 56% of financial institutions surveyed cited synthetic identities as their top fraud concern, and 40% had seen increased attack rates tied to GenAI.
Therefore, the critical key to securing a national digital identity system is to ensure that the liveness of any ‘selfie’ can be trusted. Liveness detection technology confirms whether a biometric input is from a live person − not a photo, video, deepfake, or mask − and instantly matches live selfies against captured credential images.
Advanced liveness detection systems use multilayered technologies such 3D or 4D scanning to assess factors like micro-movements, skin texture and lighting consistency. Ideally, liveness detection technology should have iBeta Quality Assurance. iBeta is a NIST NVLAP-accredited independent testing lab that certifies biometric liveness detection solutions against the ISO/IEC 30107-3 standard for Presentation Attack Detection.
With an effective and secure national digital identity system in place, there would be far-reaching benefits for the public sector and citizens alike.
It would offer huge convenience − for example, people might no longer have to go and stand in a queue for four hours at the Department of Home Affairs to get an ID book or renew a passport. They could simply do it via their mobile phones. Grant recipients could potentially be verified faster and more cost-effectively, so reducing fraud and making the payments process more efficient.
In national education, learners and students could be identified immediately, improving safety on campuses and potentially reducing fraud and cheating in exams. There are potential use cases in healthcare, social services and border control, among many others.
But these benefits can only be realised if gaps in the current plan − notably liveness detection − are closed.
Share
