As consumers increasingly turn to online banking in the wake of the COVID-19 pandemic, certain services on financial platforms are starting to see increased traffic. This has led to higher rates of fraud, as bad actors strive to exploit the crisis.
What kind of fraud? According to recent DataVisor research, account takeover attempts have increased by 20% and new account fraud increased by 40% — all since the beginning of March. And as government bodies issue stimulus packages, there’s been an increase in malicious domain registrations, which can be used to perpetrate email phishing campaigns such as emails pretending to deliver payouts. According to Google, nearly one-fifth of all phishing emails in Gmail are coronavirus-related.
The increase in fraudulent activity in the banking sector isn’t likely to ease up, as shelter-in-place orders remain active throughout May and possibly into the summer months. Consumers will continue to leverage online banking apps, opening the door for fraudsters to login and cash out.
Financial Fraudsters Have Many Vectors
Stopping modern fraudsters from attacking financial institutions isn’t easy — and it requires increasingly advanced techniques. That’s because fraudsters are adept at evading detection by blending in with the normal activities of legitimate users. For example, they may randomize the timing of their attempts in order to avoid velocity-based bot detectors. They may use fake contact information and scripts to generate realistic looking email addresses or use emulators and jailbroken mobile devices to create the appearance of multiple independent customer accounts.
To make matters worse, increased reliance on mobile banking apps broadens the threat landscape and provides a vastly expanded attack surface for bad actors to initiate these malicious activities. Data must be collected and analyzed holistically at the source to stop fraud before it infiltrates the data network.
Today’s fraudsters are quick to evolve their tactics, rendering traditional fraud detection methods — many of which use statistical analysis based on existing datasets — ineffective. What’s needed is an approach that can provide early detection of both known and unknown threats and enable fraud and risk teams in financial institutions to stop fraud at the gate.
Advanced Machine Learning: The Key to Secure Online Banking
Over the past several years, fraud detection has employed supervised machine learning (ML). In this type of ML model, data from past transactions is labeled as fraud or not fraud, then the model learns the patterns and analyzes new data based on what it knows to identify anomalies. The problem is that new types of fraud attacks emerge all the time, and models trained on past data may not be able to spot them. Additionally, they can result in a high number of false positives — in the form of a declined ATM or credit card, or blocked access on a mobile app. Although the organization is protected from potential threat, the customer experience suffers.
Advanced models that leverage unsupervised machine learning (UML) techniques are able to identify potentially fraudulent behavior by spotting unusual patterns in the data, even in the absence of labeled transaction data. In addition to anomaly detection, UML uses clustering and graph analysis techniques to uncover relationships between input data. In this way, they can detect potential threats in real time and help stop an attack before it wreaks havoc on customer accounts. UML is especially effective for discovering new and unknown patterns, which is useful for thwarting today’s sophisticated fraudsters.
Additionally, UML models dramatically reduce false positives because they are more precise and accurate than traditional ML models. This helps remove friction from the customer experience.
Safe, Frictionless Banking During the Pandemic and Beyond
The trend toward online banking via browser and mobile apps will continue to gain momentum, as Americans continue to become accustomed to interacting with brands across many industries — banking, retail, healthcare and more — from home. Financial institutions that implement proactive, early detection strategies and techniques for stopping financial fraud can ensure safe banking and deliver a seamless, friction-free customer experience that gives them a competitive edge.
