[ad_1]
Black hats or crackers are hackers with malicious intentions. — Image by Volker Agueras Gäng CC BY 2.0.
The company Jumio has unveiled a new fraud-fighting technology that uses AI-driven predictive analytics to identify fraud patterns across the globe. This technology has identified that 25 percent of fraud is interconnected.
The technology is termed Jumio 360° Fraud Analytics and it uses AI to predict the likelihood of fraud whenever a user goes through an identification verification process.
The system looks beyond simple linkages, such as flagging someone because they are connected to a known fraudster. Instead, the algorithm looks at billions of data points across a cross-industry network to identify patterns based on behavioural similarities and other indicators.
This is fraud either perpetrated by fraud rings or by individuals using the same information or credentials to open new accounts online (such as banking sites, ecommerce platforms, sharing economy sites and others).
Through such measures, a single organized crime operation or fraud ring can cause damage into the hundreds of millions of dollars. By using new technologies, fraudsters are becoming ever-more sophisticated.
The Jumio 360° Fraud Analytics seeks to stop fraud. This is achieved by moving beyond simple linkages, such as flagging someone because they are connected to a known fraudster. Instead, the technology deploys graph database technology with a layer of machine learning. This makes it easy to see connections between people, documents, and devices.
Furthermore, the combination of graph database technology and machine learning classifies identity transactions into clusters based on behavioural similarities, which is especially powerful for identifying fraud rings.
Once the AI has grouped identity transactions into clusters it determines the fraud risk of each cluster. This provides a multi-dimensional view of each transaction and the cross-customer ecosystem as a whole. Inferences are then drawn using predictive analytics.
When an individual begins to go through an identification verification process, the AI compares the information presented to different data clusters and generates a predictive fraud score that can be used to automatically reject the transaction if it exceeds a certain threshold.
In trials, the fraud detection rate has risen by over 30 percent without increasing the false rejection rate. The system allows for user assessment should there be a dispute. Dashboards provide transparency and help you visualize connected data.
[ad_2]
Source link
