FiVerity, a threat intelligence provider, recently added its new machine learning-based cyber fraud detection solution, Collaborative AI Platform. With the growing threat of SIF, or synthetic identity fraud, the ascent of this platform has urged government agencies, banks and credit unions alike to address the issue and adopt a solution.

Setting Collaborative AI Apart

Combining both fake and real data to develop false identities, the Federal Reserve estimated that SIF has cost the financial sector more than $20 billion, including 20% of all credit losses. This particular crime is both especially difficult to detect and easy to recreate on a grand scale. Fraudsters often generate multiple fraudulent accounts and loan operations at a time, beating out banks’ prevention efforts. 

FiVerty’s new feature, unlike many of its kind in the market, touts a machine learning approach to mollifying synthetic fraud. Typically, companies will use PII—consumer personally identifiable information—to share information about this kind of crime. For Collaborative AI, financial service companies will submit customer account data to the platform, churning out reports that instantly score the data with risk level. If the risk is high, the company will reject the account. The machine learning aspect of the technology is pattern recognition; the more applications it views, the more improved tactics for fraud detection. While its knowledge base is small being only in its initial pilot phase, approximately 20 banks and credit unions are now testing its reliability. If these tests succeed, the feature will be opened to more financial service and government agency usage.

FiVeity is Not Alone

Collaborative AI is being researched in conjunction with the National Cyber-Forensics and Training Alliance (NCFTA). A public-private partnership, NCFTA and FiVerity jointly published a security report titled “Advisory on Synthetic Identities used by cyber-criminals to infiltrate financial institutions,” outlining the specialized tactics fraudsters have employed of late.  

“The FiVerity Platform seeks to bring several great weapons to this fight,” says Matt LaVigna, CEO of NCFTA. “It gets everyone past the old PII roadblock, and it leverages but greatly improves upon legacy rules-based methodologies. We are looking forward to working with FiVerity as a new research partner in the fight against today’s well-funded and fast-moving bad actors.”

About the science of the solution itself, Greg Woolf, founder and CEO of FiVerity, says “We’re very pleased to see our machine learning solutions detecting more than 35% of previously undetected SIF on a stand-alone basis at financial institutions, which can only improve through industry collaboration. While SIF is a great example of the type of complex and difficult problems that collaborative machine learning can solve, it’s only one of many cyber fraud challenges facing financial institutions.” He finishes, “We’re just getting started on solving them.”

ABOUT THE WRITER

Olivia Baker is a Tech Innovation Fellow at Identity Review from Columbia University, where she writes on tech policy and national digital identity technologies.