<img height="1" width="1" src="https://www.facebook.com/tr?id=316029722675592&amp;ev=PageView &amp;noscript=1">

At the Know Identity Conference held in March 2018, Socure’s founder and chief strategy officer, Sunil Madhu, shared his perspectives on the use of machine learning (ML) and artificial intelligence (AI) for identity verification. Sunil discussed his unique insights, gleaned from Socure’s market-leading predictive analytics platform, on a panel with the CEOs of ThreatMetrix and Emailage. Sunil’s views were often not shared by his fellow panelists, leaders of traditional fraud prevention companies that rely primarily on human-developed rules-based systems. Sunil argued that human-developed rules hinder effective machine learning, and in order to properly address fraudsters machines must be allowed to go beyond human intuition.

The discussion covered a variety of points and perspectives about the ability of ML and AI to impact the fraud epidemic, and the whole panel is worth a watch.

Topics discussed include:

  • Whether the hype cycle surrounding AI and ML is justified, and how ML has evolved since the 1980s to become a core component of many systems that power the planet today.
  • How to minimize false positives in fraud detection by using advanced modeling techniques.
  • The debate over tokenization and nonreversible data. According to Sunil, tokenization is a fallacy that’s marketed as a security panacea, but it only addresses authentication, and not authorization.
  • Stolen identities vs. fake identities. How difficult is it for fraudsters to obtain each of these, and how can security measures be better about detecting them? Sunil explains the importance of social networks to verifying true online identities; one can up the ante in fraud detection by demanding proof of a social cohort, rendering stolen data essentially useless.
  • How ecommerce security has changed from web to mobile and why the mobile-first generation is driving the economy—and how to think about fraud prevention. The combination of enrollment, native fingerprinting of a device, and biometric tokenization—paired with an authentication process installed by the device manufacturer—is the wave of the fraud prevention future and will eliminate the need for email, phone number, and other outdated verification tools.
  • The effectiveness of rules-based approaches and their inherent biases (and a debate over whether bias can be good or bad). What happens when you have 1,000 rules and fraudsters can circumnavigate them anyway?

Watch the full panel debate to get insight into how we’re thinking about the future of identity verification and what makes Socure’s approach unique in the marketplace.

Topics: Machine Learning, KBA

Socure

Socure

Socure is the leader in high-assurance digital identity verification technology. Its predictive analytics platform applies artificial intelligence and machine-learning techniques with trusted online/offline data intelligence from email, phone, address, IP, social media and the broader Internet to authenticate identities in real-time. Socure powers financial inclusion, increasing acceptance as much as 40 percent for millennials and other thin-file consumers. It also reduces fraud for online new account opening by up to 90 percent, lowers manually reviewed knowledge-based authentication (KBA) rates by as much as 80 percent, and automates Customer Identification Program (CIP), Know Your Customer (KYC) and anti-money laundering (AML) compliance initiatives. Socure was founded in 2012 and is based in New York City.