Experts have welcomed a Financial Action Task Force report highlighting new technologies that can help fight financial crime. The task force emphasized how advances — such as customer due diligence tools, artificial intelligence and blockchain — can make anti-money laundering and counter-terrorist financing (AML/CFT) efforts faster, cheaper, and more effective. There are still barriers to their adoption, but these can be overcome, said the report.

Emerging technologies allow financial firms to identify and manage crime risks more quickly and effectively, and help data collection, processing and analysis. They can also support firms’ onboarding practices, relationships with regulators, and accountability.

More accurate identification systems, monitoring, recording and information sharing could also boost crime-fighting efforts, which are behind the curve in some firms, said the FATF. It highlighted the potential of solutions based on AI, and its subsets of machine learning and natural language processing, to help identify and respond to risks.

Dev Odedra, AML expert and director at Minerva Stratagem Consulting, said: “As anti-financial crime technology advances, firms should take careful note of this report. As criminals use technological advances, so should efforts to stop them. However, it may be easy to get caught up in the increasing wave of new technologies, so firms must first think about what they want the technology to achieve and whether it will be effective rather than implementing then trying to find a meaningful use.”

What are the New Technologies?

The report featured research with companies and financial technology experts who said the solutions with the most potential to improve AML/CFT effectiveness are AI, application programming interfaces (APIs), and customer due diligence (CDD) tools. They also mentioned distributed ledger technologies, such as blockchain, but these had less adoption among respondents.

Machine learning offers the greatest advantage as it can learn from existing systems, reducing the need for manual monitoring inputs. This decreases false positives and helps filter cases that need more investigation.

In remote onboarding and authentication, institutions can use AI, biometrics, and liveness detection to perform microexpression analysis, anti-spoofing checks, fake image detection, and human face attribute analysis. AI can also monitor business relationships, behaviors and transactions.

Machine learning algorithms can group customers by behavior, enabling risk-based controls and quicker or real-time analysis. They can identify activity patterns and any need for enhanced due diligence.

Odedra said: “Technology can do repetitive tasks faster and more accurately than humans, from entering basic know your customer (KYC) data to searching volumes of unstructured data. This allows humans to concentrate on crime risk decisions that require their input.”

“Firms increasingly use AI in transaction monitoring to find activity of concern and previously unidentified risks. It also helps reduce false positives, which saves a lot of wasted effort.”

Odedra said AI combined with human thinking would become much more effective in spotting suspicious activity compared to previous efforts. 

“Also, adverse media searches combine with AI to reduce repetition and manual effort,” he added. “Open source intelligence searches conducted by human investigators, should also not be underestimated in AML investigations. Meanwhile, blockchain analytics software has proved invaluable in tracking and tracing funds connected to cryptocurrency-hacks and ransomware attacks.”

Benefits of Artificial Intelligence

The FATF report said multinational financial institutions and FinTechs have led the demand for such technologies. Most agree they improve AML/CFT effectiveness, risk management, speed, flexibility and governance.

For example, one multinational institution is building a dynamic risk assessment tool, enabling it to:

  • use data with greater depth and richness, updated dynamically to reflect the latest investigative insights
  • identify financial crime risk faster with fewer unproductive alerts
  • assess customer risk more accurately.

This tool uses the cloud to centralize and process data at scale; and new techniques, such as machine learning, to identify financial crime risk by:

  • integrating existing knowledge on financial crime types and suspicious activities
  • looking at transactional and social links to other entities with suspicious or adverse characteristics
  • capturing abnormal behavior compared to peer groups and previous actions.

The report also looks at how financial institutions in Brazil are using machine learning, analytical tools and integrated databases in monitoring and CDD processes. This is helping them identify new ML/TF risks, and increase speed of analysis and assertiveness of alerts, among other efficiency gains.

Technology Challenges

The main challenges in adopting technologies include operational constraints, legacy AML/CFT compliance systems, and traditional regulatory frameworks.

Difficulties with explaining and interpreting digital solutions stem from the limited expertise and low awareness of technologies among AML professionals. Increased communication and cooperation between the public and private sectors will be vital in overcoming these challenges.

Risk-targeted CDD efforts can be inaccurate and irrelevant. They often rely heavily on human input and defensive, box-ticking approaches, rather than a more risk-based approach. Legacy systems are static, obstructing a real-time view, and do not allow more detailed analysis through large-scale data processing.

Such a defensive approach is inefficient, costly and burdensome, and does not reflect the real ML/TF threats to institutions. It undermines a genuine risk-based approach by focusing too much on low-risk situations and depleting attention on new or emerging risks. The result is money laundering and terrorist financing going undiscovered; unnecessary costs and friction to customers; and financial exclusion.

Greater ability to collect and process data, and share it among stakeholders, could promote a more dynamic risk-based approach. Machine learning and other AI-based tools allow for faster, more accurate data analysis, at an optimized cost, which may solve these issues and enable financial inclusion.

The report also highlighted the importance of digital ID, which has been widely adopted in many countries. Electronic ID and verification are among the “most mature and instantly useful technologies in AML”; and are often associated with good practice.

Enhanced technologies for client screening and matching can also improve compliance because reliance on out-of-date and regionally irrelevant sanctions and politically exposed persons (PEP) lists is an area for improvement.

No Time for Coasting

Odedra summed up the implications for AML professionals as follows: “Previously it may have been easy for some to coast in their jobs - I often saw them only keep an eye out for regulatory updates, which are infrequent. Today, financial crimes and the technological tools available to criminals are increasing.”

“But those looking to spot and stop criminals are also using technological advances. AML professionals should not just stay alert to regulatory changes but, more importantly, the ways and technologies available to financial criminals, and the technologies available to help them stop crime.”