PARTNER

Using data to power the next generation of fraud detection

By Cloudera

Cloudera discusses how data can help governments detect fraud and financial crime.

While Covid-19 relief funds were created to help the needy, a thieving hand was operating in the shadows. In the United States, nearly US$100 billion was stolen from Covid-19 relief programmes, according to the U.S. Secret Service.

Fraud and financial crime is not just a pandemic concern, however. US$42 billion was lost to fraud over a 24 month period, according to the PwC Global Economic Crime and Fraud Survey 2020.
Cloudera discusses the challenges governments face in detecting fraud and financial crime, and how using data can help.

Weeding out fraud with data analytics

Machine learning and AI can help governments detect suspicious behaviour, identify fraudulent threats and shut them down.

For example, the US’s Internal Revenue Service (IRS) used Cloudera’s data and analytics platform to detect fraud and financial crime. Cloudera’s platform created graphs that identified patterns in financial transactions among individuals and organisations. Next, AI and machine learning helped to detect anomalies in these patterns that signal potential fraud.

IRS used to take up to months to compile and analyse its data. Now, it takes just minutes with Cloudera’s platform, which is supported by NVIDIA’s processing power.

Down under, the Australian Taxation Office uses data to detect fraud by comparing data from its systems with information from its partners. This data is then used to identify suspicious activities.

Meanwhile, Indonesia’s tax department, Direktorat Jenderal Pajak, examines taxpayers’ behavioural data to predict fraud and reduce late tax payments. The tax department worked with the Behavioural Insights Team resulting in Indonesians paying USD13.53 million extra in taxes by the deadline in 2017.

How data powers next-generation fraud detection

Governments need the right tools and technologies to effectively combat fraud with data.
For example, data gathered from an individual’s social media, location information, obituaries or repair costs helps banks quickly identify suspicious insurance claims or applications.

Governments can rely on data to identify taxpayers who are more likely to commit fraud and flag them for auditing, allowing for a more efficient distribution of resources. Data tools can identify these individuals by analysing their police records, property records, and cross-referencing them with other tax department databases.

For example, Singapore’s Accountant-General’s Department uses bots to comb through their data. These bots process 100,000 records in 10 minutes, allowing the department to quickly identify discrepancies in the data and conduct more thorough checks.

Advanced tech such as AI and bots can help governments in their fight against fraud. However, it is important for governments to do this responsibly by collecting data under the proper governance and compliance policies.

Challenges governments face in detecting fraud and financial crime

Many cases of fraud and abuse go undetected due to insufficient risk assessment measures, obsolete technologies, and inadequate or lack of data strategies.

Without the right data strategies, governments cannot use data effectively. This prevents them from accessing historical data to proactively prevent fraud and financial crime. Lack of access to real-time data also hinders their reaction time when fraud and financial crime do happen.

Another huge challenge governments face is the sheer prevalence and varied nature of fraud and financial crimes. Fraud is the second most disruptive event for the government and public sector, according to PwC’s Global Economic Crime and Fraud Survey 2020.

These crimes can originate internally, externally, or even as a result of collusion. It is not uncommon to hear of crimes perpetrated by business partners or even by employees.

Multiple potential avenues of risk makes it difficult for governments to determine which data to use, how they manage data, and what they can do with the numerous analytics methods available to them.

“A single tool or technology on its own will not amount to an anti-fraud programme,” according to PwC. But having a variety of tech tools and policies available can help governments stay on top of fraud and financial crimes.