Artificial intelligence and machine learning essential in the fight against fraud
A guest contribution by Anders la Cour , co-founder and Chief Executive Officer of Banking Circle
Digitization has picked up considerable speed in recent years. During the pandemic, this change in the financial industry accelerated dramatically. At the same time, the increased digital exchange also represents a gateway for fraudsters. However, the number of money laundering cases was already on the rise before financial service providers were forced to move their operations to remote workplaces.
Already in the first six months of 2020, money laundering fines worldwide reached more than $700 million. That’s nearly double the 2019 figure ($444 million)[i]. While the fines are small relative to industry turnover, the damage caused by lost customer confidence and business disruption is significant. To mitigate this damage, banks spend an average of $48 million each year on know-your-customer (KYC) and anti-money laundering processes.[ii] off. This high level of spending explains why financial institutions are consolidating their strategies to reduce overall risk – and have restricted offerings to certain sectors.
However, excluding a market or a sector is not an optimal solution. For example, many financial institutions have adopted artificial intelligence (AI)-based approaches to combat the rise of money laundering. However, this also brings some challenges:
As part of a study, Banking Circle spoke to 300 senior decision makers in European banks in November 2020. At the same time, many were convinced that AI implementation to date has been far too inconsistent and may be jeopardizing business goals.
In addition, resilient implementations are hampered by increasingly tight IT budgets. Nevertheless, respondents believe that AI and machine learning (ML) will be essential in the fight against money laundering in the future.
AI-based approaches on the rise
Compliance is often seen as a necessary evil. However, compliance can be leveraged to drive business and significantly improve efficiency. Traditional rules-based processes capture only one element of a transaction, resulting in false positive rates of 97 to 99 percent. AI and ML significantly improve the precision of the rules used in previous automated anti-money laundering processes. This is because they provide a number of indicators that point to possible risks. This reduces false alarms while meeting compliance criteria and successfully combating financial crime. In addition, the workload is reduced, freeing up resources among employees. These can in turn be used to focus on other areas, such as expanding customer relationships.
By now, many financial institutions have started implementing AI-based approaches. This is to combat the rise in money laundering. Many banks want processes that automatically apply machine learning techniques to data collected throughout the transaction chain. Up to now, this has generally only been done for selected parts of the process.
Companies can only take full advantage of operational efficiencies if they look at the big picture. They need to start thinking holistically about the role of money laundering and compliance within digital transformation. However, it is not easy to introduce new approaches to prevent money laundering amidst the current changes in the banking industry. That’s because traditional banks are being held back by a mix of outdated IT, shrinking IT budgets and poor data quality.
Achieving success faster together
In the fight against money laundering, partnerships are the key to success. Financial service providers need to consider both domestic and international collaborations to share data and approaches to combat increasingly sophisticated international criminal organizations.
“In the fight against money laundering, partnerships are the key to success.”
Merging data is an essential element for successful AI and ML. However, it requires that the data are clean, well labelled and from reliable sources. Finally, this data must be properly managed and interpreted. Nearly one in four (24 percent) respondents said that poor data quality is a major problem for the success of their IT strategy. They estimate that up to 15 percent of real-time transactions are blocked due to bad data about recipients or initiators of transactions.
For anti-money laundering processes to be efficient and effective, data quality must improve dramatically. Cross-industry collaboration is the best way to bring about this change.
The white paper “Better by design? Re-thinking AML for a digital age” with additional information is available for download here.
For further reading:
IT Governance, Risk and Compliance – from a must to added value
To continue listening:
Money Laundering – FinTech Podcast #176
Source:
[i] Source: https://www.ft.com/content/a547e6ed-5a2e-48c4-bbee-febbf975e4af
[ii] Source: https://home.kpmg/mc/en/home/insights/2019/03/combating-financial-crime-fs.html
