GBG has won the highly sought-after award for Best AI or Machine Learning Innovation of the Year at the Asia Risk Awards 2021.
This award win shines a light on GBG’s market leading machine learning (ML) capabilities, which uses intelligent algorithms to uncover inconsistencies in behaviours and activities, analyse them, and alert systems operators about anomalies.
Unlike traditional rule-based systems, GBG’s flagship financial crime solution – the Digital Risk Management and Intelligence platform – utilises machine learning to detect invisible correlations in data in real time, comparing with historical sets of personal data, to uncover new fraud instances as quickly as possible — whilst concurrently ruling out the possibility of the anomaly being caused by accidental human error, and rather, as the result of a bad actor.
Machine learning is at the heart of innovative financial crime detection
GBG’s win of the best AI or Machine Learning Innovation of the Year comes at a time when financial institutions are increasingly considering the effectiveness of this technology to reduce fraudulent activity. According to a recent GBG survey conducted in April 2021, more than one third (36%) of financial institutions in South-East Asia are exploring machine learning to help combat modern financial crime.
Today, Southeast Asia’s first fraud bureau, CTOS IDguard, which is powered by GBG’s fraud and financial crime platform, is currently deploying GBG Machine Learning and is seeing an uplift in fraud detection of up to 30% for credit card applications. Another Tier 1 global bank has also tested its machine learning technology to see a 22% reduction in false positives for credit card applications.
The evolution of modern financial crime is presenting financial institutions with not only a growing volume of fraud instances to manage, but of greater concern is the creativity and sophistication in which fraud attacks are executed.
Consequently, GBG Machine Learning has been designed with supervised and unsupervised learning and custom models to meet varied industry needs. Supervised learning uses three proven algorithms for fraud detection. The Neural Network, Gradient Boost, and Random Forest algorithms learn patterns from past or marked fraud records and new emerging fraud attempts. Unsupervised learning uses Isolation Forest, detecting outlier patterns by using available data without past or marked fraud records. Custom modelling with Python allows the import of pre-trained models developed using other algorithms on a third-party toolkit.
GBG’s custom modelling option allows users to employ different algorithms of their own choosing, via a Python notebook-like interface, enabling financial institutions and enterprises to develop and implement solutions tailored to the efficiency and accuracy standards required by themselves and their customers.
This best in class offering, which integrates with GBG’s award-winning Digital Risk Management and Intelligence platform and GBG Intelligence Center, ensures businesses have the end-to-end fraud, risk, compliance, and identity management solution required to mitigate the impacts of some of their biggest fraud-related challenges, including high false positive rates, high fraud alerts, and missed frauds.
This was specifically recognised by Asia Risk as being a key reason for GBG’s award, with Blake Evans-Pritchard, bureau chief for Asia Risk, commenting, “The judges were most impressed with the performance of GBG’s artificial intelligence and machine learning to help reduce false positives and improve fraud detection of emerging typologies.”
GBG Machine Learning is already helping financial institutions across Asia Pacific with a range of challenges. For example, there are instances of account takeover cases where a fraudster is obtaining credentials of a legitimate customer through a data breach, then performing daily transfers just below the daily limit. As these behaviours are atypical of the original consumer, GBG Machine Learning can recognise and raise an alert of the anomaly.
Another example involves banking customers using machine learning to detect and stop money exchanges used to fund child exploitation and wildlife trafficking. In these cases, machine learning is used to reduce false positives, using rules based on parameters such as country locations, or money transferred of a certain amount or frequency. In these situations, machine learning has also helped to increase banks’ accuracy of identifying and locating money laundering activities, further enabling banks to predict potential future cases.
To get a demo of GBG’s end to end financial crime solution and machine learning, click here
To find out more about GBG Digital Risk Management & Intelligence platform, click here
To find out more about GBG Machine Learning, click here
To read the press release of GBG Machine Learning / AI Innovation of the year, click here