menu-close
search-icon
banner

All posts by Nithin Gangadharan

Artificial Intelligence and Machine Learning – the key to combating Identity Fraud

In an increasingly digitally connected world, identity fraud is a growing problem. According to the Communications Fraud Control Association’s (CFCA) annual fraud survey, identity fraud took the top spot as the number one fraud method present globally and at individual companies. Also featuring as the first among the top ten fraud methods, identity fraud in telecommunication during the subscription process (subscription fraud) resulted in costs of $2.03 billion.1

Identity theft, where the use of fabricated identities at point of sale, enables the fraudulent use of telecom services and/or perpetuates subsequent fraudulent activities, can result in serious implications in the present age. With highly interconnected 4G and soon to be launched 5G networks enabling not just value-added services but also financial services like mobile payments and banking, it opens up access like never before.  Identity theft can work as an entry point for myriad types of fraud or even terrorism. With access to secondary authorizations (PIN code verification), subscription fraud can be used for any number of illegal activities.

This has facilitated an urgent need for a proactive and dynamic response to fostering digital identity security.

 

Securing against Identity Fraud

The answer to fighting fraud lies in detecting data anomalies in real time. For example, to tackle telecom fraud, the Technology Research Institute (TRI) has stated that, real-time point-of-sale identity verification services are an invaluable aid to stopping fraudsters from exploiting identity theft.2 Historically, rules have always been in the system. But in the increasingly connected world, effective fraud coverage is only possible with a combination of rules and applied Artificial Intelligence (AI) and Machine Learning (ML) technologies.

With AI and ML technologies, companies are able to detect data anomalies in real-time and make decisions based on information as it happens, empowering them to anticipate and take proactive action. For instance, AI/ML techniques can use facial recognition technology to identify high risk by making checks against blacklists. ML can augment traditional rule-based systems to develop and train algorithms to determine the characteristics of traffic and identify anomalies that could end up being fraud.

Furthermore, the immense amounts of unsecured data flowing in from connected devices onto operator networks can be secured only with AI and ML. AI technologies are equipped with the capacity to scale up efforts and enable fraud detection at a massive scale by handling the management of millions of customer or network data points.

As networks continue to expand and new fraud schemes continue to evolve, a combination of rules and applied AI/ML models will serve as the most effective way in combating identity fraud.

If you are interested to learn how AI /ML techniques can help you combat Identity Fraud

DOWNLOAD WEBINAR RECORDING NOW!

 

1 . http://v2.itweb.co.za/whitepaper/Amdocs_LINKED_2017_CFCA_Global_Fraud_Loss_Survey.pdf

2 . http://technology-research.com/products/fraudmgt/telecom_fraud_management_executive_summary.pdf?_sm_au_=iTHSrnwsn4HFwT3q

Why Artificial Intelligence Powered Fraud Management

Artificial Intelligence (AI) is not new and it has been around for decades. However, with the advent of big data and distributed computing that is available today, it is possible to realize the true potential of AI. From what started as an interesting story line in SCI-FI movies to programs like Alpha-Go which has been beating humans, AI has been evolving. AI also has branched out into multiple sub categories such as Machine Learning, Deep Learning, Re-enforcement learning etc.
FM-1

FM-2
An effective Fraud Management (FM) strategy includes 3 important pillars: Detect, Investigate & Protect. We believe AI can positively influence all the 3 pillars of fraud management, from reducing false positives to helping in mining root cause analysis to creating enhanced customer experience in protection.

In this post I would like to look at the starting pillar of the Fraud Management strategy – “Detection” and look at AI’s influence in this very important step. A traditional approach to Fraud detection has been through Rule Engines which could be:

  • If-Else Conditions
  • Thresholds
  • Expressions
  • Evaluating Data Patterns
These are widely known as deterministic solutions where an event triggers an action. The biggest pros and cons with this approach is that human intervention is needed to feed the logic.

For eg: for a threshold based detection humans have to feed the rule engine that count of records above a certain threshold is suspicious.

Following diagrams shows how this looks like

rule-engine

After looking at the diagram above an important question arises, should this threshold value be a straight line or can it bend based on how data behaves. Now there are ways for rule engine to behave like mentioned in the diagram,

variable-threshold

for eg, instead of having a single rule lets have multiple rules

  • Per Customer Category
  • Per Destination
  • Per Age of Customers

And multiply that with other dimensions in data which are

  • Phone Number
  • Caller Number
  • Called Number
  • Country Code

And multiple that with other set of measures per dimensions

  • Count
  • Duration
  • Value

And throw an additional billion volumes at the datasets

Quickly FM teams ends up with something like this
AI Blog1
But what they wanted or dreamt was this
AI Blog2

Now I am not saying FM teams are not skilled enough to fly, but a fraud team in a modern Digital Service provider should be more focused on other important factors.

machine-learning
So, let’s look at how a very evolved class of Artificial Intelligence known as Machine Learning looks at this problem statement. Rather than humans feeding domain information or thresholds, Machine Learning Algorithms mine data from historic fraudulent behaviors and create models. These models are then used to evaluate real production datasets to score whether they certain activity is fraud or not. An advantage is that these models are very good at looking the datasets from multiple dimensions and measures at the same time and concluding whether event is fraud or not.

This approach thereby helps in achieving multiple KPI’s of fraud management teams there by increasing efficiency.

  • Higher Accuracy – Because AI can learn and adapt to Business scenarios faster, AI can significantly increase True Positive ratio
  • Reduced time to detect – How fast a fraud event can be detected
  • Self-Learning – How over a period changing business scenarios and seasonality in data can be adopted to Fraud detection
  • Fraud Intelligence– How customer or any other entity behaviors can be learnt and categorized for better fraud detection
  • Proactiveness – Ability to mine for unknown patterns not seen in the data earlier
FM-4

Application of Artificial Intelligence has its own significant challenges and requires a new frame of thought, however looking at the Data Tsunami that has hit the fraud management teams, it looks an AI pro approach would only help Fraud Management teams to scale further.

Why Telcos could never overcome Simbox Fraud since a decade Now

Simbox, Bypass Fraud/ Or Interconnect bypass Fraud has been one of the fastest growing Fraud Types In recent few years.  As per 2017 Global Fraud Loss Survey by CFCA, Global Fraud Loss Estimate stands at $29.2 Billion (USD) annually which is 1.27% of global telecom revenues.

global bypass

Source CFCA Survey Results

Simbox Fraud / Bypass Fraud has been a significant fraud issue for more than a decade now. CFCA survey results across 2009 till 2017 clearly shows an increase of more than 100%  in Bypass fraud since 2013. In this blog, we shall discuss about factors that has contributed to this continuous increase in Bypass Fraud and reasons, operators have not been able to effectively mitigate Bypass Fraud.

Factors for continuous Increase in Bypass / Simbox Fraud:

  • Reduced barrier for entry

Buying and operating SIMBoxs has never been easier with online stores, e-commerce websites,courses, forums and instant support availability.  This has led to an increased spread of VOIP based startups and subsequent increase in bypass fraud. VOIP based calling apps have also made customer acquisition easy by making them  easily available on  AppStore for Android & IOS users. For instance, a recent news from India covered the similar trend wherein those who wanted to make international calls from Gulf countries has to download an app called ‘dial to India’ Once this app is downloaded, they get a password for monthly subscriptions. The person sitting abroad will just dial the number in India, the call will bypass the VSNL gate and will directly route through the SIM box and will get connected from there. Read More

Few more such examples as below:

Illegal phone exchanges thriving on SIM boxes

VOIP exchanges used by ISI busted in Andhra Pradesh, India

  • Competitive Landscape

Reduced margins on international traffic has resulted in wholesale traffic being mixed with internal traffic. Wholesale providers have also been increasingly offering non-CLI based options which could potentially end up in Grey routes. This fierce competition had led to increase in bypass traffic particularly in countries with higher landing costs.

Reasons Operators have not been able to effectively mitigate Bypass Fraud:

  • Advancement in Sim-server Technology

Simbox have evolved from being a simple single box setup to a complex modular architecture. This architecture allows fraudsters to maintain all the simcards in a single place and using Antenna modules and multiplexers, fraudsters are able to distribute their operations in the market. In fact, Latest Simservers also comes with inbuilt anti-fraud detection solutions allowing fraudsters to  distribute his operations in multiple locations. This makes fraud detection very complex as fraud management teams have to device multiple strategies to beat fraudsters at their game.

  • Regulatory Changes

Regulatory changes in certain markets have fueled increase in traffic for Bypass. Recent changes of regulations in European Union has also resulted in traffic with E.U been heavily being differentiated in price from traffic outside E.U thereby causing significant increase in Bypass traffic.

  • Raising Concerns in Simcard Sales

Increased pressure to maintain sales and activation of new connections have resulted in dealers colluding with Bypass fraudsters. Bypass operations requires lot of sims to be activated in bulk and lack of effective subscriber acquisition controls have led to fraudsters taking advantage of it.

Fraud Management teams further have an uphill task in the Bypass fraud space as new technologies such as virtual sim’s would only increase the impacts on bypass of international traffic. It is hence important that they adopt a comprehensive fraud detection methodology to fight simbox frauds.

Why inline fraud management is necessary to combat POS Frauds?

Over the years Fraud Management teams have been more reactive in nature and Frauds were usually detected once the event has occurred and the losses are incurred. With the revenues from the traditional services such as Voice & SMS crumbling down and operators looking for newer avenues to sell products and services, it is essential that the wafer thin margins are well protected. This has also forced the Fraud Management teams to be more proactive in nature.

One of the key focuses of being proactive is to prevent Fraudsters from entering the network. Fraud protection at point of sales stage has got an increased focus in the recent years. CFCA Fraud survey report of 2011 estimates global Point of Sale fraud at almost $5.5 billion. One other startling fact reveals that UK Mobile Operators lose £136M annually to fraudulent transactions at Point of Sales.

Some of the traditional challenges at Point of Sale include Subscription Fraud and loss due to Handsets being bundled to subscriptions. However with 4G and data heavy environments coming into play there would be more products, services and content sold over various sales channels such as Web Stores and Retail Centers.

To meet such requirements Fraud Management systems have evolved from being just Near Real Time systems to Real Time Inline systems which can integrate to various Sales Channels. On receiving the orders the system evaluates various order attributes such as Location, Subscriber demographics, Type of order, Type of Retailer and evaluates whether the order is genuine or not.

If an order is found suspicious, the request can be routed through various workflows within the system such as generating an alert for analyst approval or declining the order directly. Fraud Management system should be flexible enough to add new rules and workflows as the

new products are launched by the product teams. Behavioral profiling can be also used to track and monitor suspicious behavior patterns on the orders that have been placed.

Benefits of proactive and inline fraud management is unchallenged. Banking & Payments industry has been practicing these for a long time. By configuring and fine tuning the rules over a period of time it would be possible to detect and prevent majority of Fraud events. One of the recent implementations of Inline Fraud Management for Subscriber Acquisition helped save the operator nearly 2 Million $ in losses and it was observed that Fraud Hit ratio was about 60%. It was also noted that the Fraud Hit over the weekends where as high as 100%. However care has to be taken during such an implementation that only confirmed fraudulent orders are directly rejected by the system as any rejection of genuine orders can give a direct blow to Customer Experience.

To download infographic, click on the link: Inline Infographic

Get Started with Subex