/artificial intelligence

News and resources on artificial intelligence systems, innovations and initiatives worldwide.

Money 20/20 Asia Q&A: Onfido’s Husayn Kassai on fixing identity services

Husayn Kassai, CEO and co-founder of Onfido, spoke to Finextra Research at Money 20/20 Asia in Singapore about the broken identity verification process, the role of machine learning in this space and the benefits of a pan-Asian KYC plan for identity services.

  1 Be the first to comment

Money 20/20 Asia Q&A: Onfido’s Husayn Kassai on fixing identity services

Editorial

This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.

How has your background encouraged you to take an interest in identity services?

The identity verification process has always been broken and it is getting worse and this is why we founded Onfido in 2012. When I moved from Iran to the UK at the age of 10, it took my parents months to open a bank account and rent a home in their own name and this was because they were not on the credit bureau database. As I grew up I realised what a waste of time the credit bureau model was because it excluded half the world’s population who are either unbanked or underbanked and therefore, not on this database. For those who are on the credit bureau database, like you and I, our information is also on the dark web and available for anyone who wants it. Our plan is to remedy this with a more secure service and while we are starting with financial services, we believe that this will be transferable to the rental car industry, hotels and even voting. We are only scratching the surface but that is how it starts.

What role can machine learning play in the advancement of identity verification services?

What used to happen, or still does to some extent, is that customers would go to a bank branch with a physical form of ID and the bank clerk would have to look at their face and compare it to the image they see on the passport or driver’s license, determining if it seems fake or not. Despite facial recognition on a human level being considered better than any other form of verification, the first breakthrough in biometrics was in digital facial recognition, which was proven scientifically to outperform the human eye when conducting facial recognition at scale. What we’re doing with government IDs is using machine learning to detect faces that cannot be authenticated by the human eye. This is something that banks did not realise until around 18 months ago.

Are you working with any governments at the moment to provide identity verification services?

We are working with them indirectly, for example with organisations such as Babylon Health and the NHS, by allowing their users to access their health records. However, three quarters of our client base are financially regulated organizations, from mainstream banks to fintechs.

What benefits would an Asian KYC plan have for identity services?

There is a fundamental problem in that if you go from country to country, you have to start from scratch and it takes a long time to undergo the full journey. In France and Germany, you can open a bank account in three minutes, let alone three months, so if you can transport your digital identity, it will be 30 seconds, rather than three minutes, but the three minutes is a massive improvement from the three months issue. That is the big leap and equally here, moving from Indonesia to Malaysia to Singapore, there have been significant reductions in time for switching banks by simplifying the process but also, maintaining the security. What we want is for governments to issue their citizens with ID cards that we can verify; a good example would be Aadhar in India which has provided most of the population of the subcontinent with an identity.

Is the ID card a stepping stone to eradicating all physical forms of identification?

Being issued with an identity card from your local government is the first necessary step and we can digitise it. If you look at a centralised credit bureau model with a leaking database, we’re substituting and improving that by going half-way and verifying it. What you’re talking about – a fully digital identity – that’s absolutely the future. You enter a decentralised world where the consumer controls their data and there is no need for the institution to hold a copy of your ID, because this exponentially increases the chance of you being a victim of identity fraud.

If the future is decentralised, is identity then a use case for blockchain?

If it’s decentralised – the consumer owns their data and it is likely that there will also be a need for an encrypted app. How you do this could be distributed ledger technology, or blockchain, but this is only one option as in the Nordics, it is done successfully without the use of blockchain. The technology has some components that are beneficial, but it is not necessary: it would help but there are some aspects that need to be corrected, such as speed and governance.

Sponsored [On-Demand Webinar] AI in Banking: Building Compliant and Safe Enterprise AI at Scale

Comments: (0)

[On-Demand Webinar] Solving the KYC challenge with end-to-end processesFinextra Promoted[On-Demand Webinar] Solving the KYC challenge with end-to-end processes