fatbrain vector blog


Discussions on the direction of AI-assisted decision-making.

Artificial Intelligence

Posted February 19, 2020 by Peter B. Ritz

8 Anti-Money Laundering Predictions for 2020

(8 minute read)

Perhaps no other area in financial services is poised for big change than AML and financial crime compliance. Recent regulations and directives with an eye to stopping financial crime, maturing AI/ML technologies and federated learning approaches for sharing data across the sector (and geographies) have formed the perfect storm for banks to achieve progressive, positive results in reducing money laundering, terrorism financing, fraud and other financial crimes.

With that, here are our eight AML predictions for 2020:

Prediction #1: Non-Compliance Fines will Continue to Rise Globally

Since 2015, annual non-compliance fines from AML deficiencies have steadily risen each year, with 2019 topping out at $8.14B in fines versus 2018 at $4.27B. 2020 should this trend to continue over the short term.

Despite best efforts, banks and other money-based businesses are still slightly behind in adequately detecting and reporting suspicious behavior. The sector is getting closer, but still not close enough using the legacy rules-based systems. Large investigative backlogs still exist due to large volumes of false positives – ampable hiding spots for the latest financial crime scheme to go undetected.

Prediction#2: Two Top 20 Global Banks will Receive a $250M+ fine

2019 saw one $1B fine administered to a tier one Swiss bank, so sadly the idea of a $250M for a Top 20 Bank should almost be expected in 2020. In fact, the average fine in 2019 was $145.3M – perhaps a bolder prediction would see the average fine in 2020 being $250M.

With pressures mounting on banks to improve AML performance, the EU’s 5th Money Laundering Directive (5MLD) went into effect on January 10, 2020 will have a significant impact on non-compliance due to missed deadlines and new requirements.

5MLD introduces the need to oversee and detect activities involving virtual assets like cryptocurrencies and fine art, and monitoring prepaid cards as low at 150 euros and remote payment transactions of more than 50 euros. Companies not exposed to AML, CTF and other financial crime practises will need to ramp up quickly to be compliant under 5MLD.

Prediction #3: The Fastest Growing Segment of SAR will Originate from Crypto Currency

Since FinCEN’s May 2019 guidance on applying the Banking Secrecy Act to Crypto-based currencies, there has been a significant increase in the number of SARS involving Crypto currency and the number of filers.

According to Kenneth Blanco, director of the FinCEN, “dozens” of new entities filed their first report“, and he is encouraged by the fact that Crypto-based businesses, many of whom had never filed a SAR report prior to the May advisory guidance, are detecting and reporting suspicious activity.

Lastly, 5MLD will also fuel the growth of SARs involving Crypto currency.

Prediction #4: AI/ML Technology Finally ‘Matures’ for AML –  Explainability Arrives

It’s been eight (8) years since IBM Watson won on Jeopardy against the two best contestants in the show’s history. Since then, Artificial Intelligence (AI) and Machine Learning (ML) technologies have only gotten better, just not for the AML use-case.

First-generation AI/ML approaches are really poor at the one thing regulators (and compliance leadership) care about the most: Explainability or all the factors that led to the decision by the AI/ML system.

Complex neural networks and other AI techniques make AI-based decisions very hard to validate, test and audit. The lack of explainability is a barrier first-generation AI platforms cannot overcome. Fortunately, next-generation systems like FatBrain AI already have figured out the explainability problem. Early adopters have detected more true positives without increasing false positives, thus have seen their effectiveness and efficiency KPIs explode upwards in a short period of time. 2020 will see the transition from early adopter to mainstream use for those AL / ML technologies that can explain their decisions.

Prediction #5: AML Transformation: Effectiveness-first

With explainability delivered, AML has the potential to truly transform from a legacy rules-based automation approach to that of intelligent automation. So instead of sifting through a mountain of detected false positive transactions ( Thomson Reuters reports that 95% of false positives end up as safe, but you already knew that based on first-hand experience), effectiveness-first finds more true positives without increasing false positives.

This perhaps the real potential of an effectiveness-first strategy. More time on suspicious activities uncovers criminal innovation faster, which minimizes the damage done, inside and outside the institution, sooner.

Prediction #6: Risk Transformation: The Convergence of AML, Fraud and Other Risk

Convergence is not a new concept, as the idea has been floated around since 2011. 2020 just may be the turning point, as the maturation of technology systems for AML, fraud and other transactions risks have evolved to have about a 80% functionality overlap. Plenty of redundancy and duplication that can be optimized with an integrated, coordinated approach. Undertaking an effort to streamline the infrastructure will drive the other benefits.

Whether you’re involving with AML, fraud or other financial crimes, the core process is very similar:

  • Confirm the identity of the customer (and associated risk profile)
  • Monitor transactions for risky behavior and parties
  • Respond and investigate the transactions. Stop fraudulent or illicit activities
  • Report suspicious activities (when required) to regulators and authorities

In addition, the core operating processes for AML, Fraud and other financial crime are supported by similar data and internal operations. The biggest benefit of convergence may not be efficiencies gain, rather the effectiveness gained at preventing financial losses associated with financial crime.

Integrated data sources will dramatically improve risk visibility and detection capability. Insights that have the potential to self-evolve, making the overall system better and better each and every day. Risk decision making is improved with convergence and integration.

When the underlying systems merge into ‘one’ unified system, convergence removes duplication and unlocks efficiencies like streamlined infrastructure (hardware, software), processes simplified, and resource efficiencies.

Prediction #7: Regulators ‘Enable’ Innovation

Regulatory pressure to stop the flow of illicit funds for money laundering and financing terrorisms is still increasing. Over $12B in fines have been levied against global financial institutions over the last two years alone.

The good news is that regulators are also taking a different approach to compliance. With the objective to stop the financial crime that is detrimental to our society, regulators are balancing inspection with the push for collaboration and innovation. Many of us know that current rules-based approaches have hit their maturity, yet criminals continue to succeed in exploiting the global system. Sadly, more financial crime is occuring, not less.

With the drive to stop money laundering, terrorism financing and other financial crimes, regulators are stepping up efforts by enforcing reduced fines when non-compliance is of result of new approaches and innovation. The U.S. joint announcement in December 2018 signaled the regulators preference for innovation, with 2019 seeing the launch of the UK’s FCA Global Financial Innovation Network.

Regulators are also taking the lead to drive collaboration, openness and change within the financial sector. Under the leadership of Nick Cook (Director of Innovation), the FCA has hosted a series of AML and Financial Crime TechSprints, a FinTech ‘hackathon’ designed to collaborate on innovative approaches and technologies that can fight financial crime. These events are bringing the ecosystem together and aligning actions across the sector, paving the way for a sector-based approach to AML.

Prediction #8: Progression Towards a Sector-Based Sharing Approach for AML

With regulators signaling openness towards innovation within a financial institution, 2020 should see progressive steps towards bringing the sector together as a whole to fight financial crime.

While organizations like the Wolfsberg Group have existed for years to share best practises to combat money laundering, laws (especially data privacy laws) and technologies haven’t stayed current to enable this effort.

2020 should be a watershed year for technology being the catalyst to evolve the financial sector. Both federated learning technologies and privacy-enhancing technologies (“PET”) have the promise to enable financial institutions to share insights within the sector without violating multi-national data privacy laws like GDPR, California’s Privacy Act (CCPA), and PIPEA.

Sector-based data sharing creates the potential for meta transaction monitoring. This enables a big picture view to transaction monitoring, with each financial institution having a piece to the puzzle and share insights and typologies. With each institution having one piece to the puzzle, the sector is better positioned to detect complex money laundering schemes across geos and institutions, while allowing for sharing insights about very specific typologies, and detect specific behavior.

Not sure where your AML efforts should head next?

Validate your strategy with the latest thinking from our team of financial crime compliance executives and data scientists in a personalized and interactive 1:1 session.

Request the initial no-charge briefing today

The session is ideal for compliance and risk executives looking for the latest strategic thinking in AML, Financial Crimes Compliance and Dynamic Risk Intelligence.

Tags: , , ,

Let’s get down to business.

Learn how FatBrain will help you effectively navigate your ever-growing data maze to quickly monetize troves of transactional data with real-time decision making.

Talk to FatBrain