Rising financial crime volumes

Fraud resulted in losses of $12.5 billion for US consumers in 2024—25% more than in 2023, according to the Federal Trade Commission. This continues a troubling trend: 2023’s reported total of $10 billion was up 14% from $8.8 billion in 2022, reflecting the growing sophistication of financial crime. 

Additional dangers are making old problems more difficult. On November 13, 2024, the US Treasury Department's Financial Crimes Enforcement Network (FinCEN) issued an alert advising institutions to carefully review identification documents, pointing out increased suspicious reports of deepfakes, including attempts to use fraudulent documents to circumvent identity verification.


Regulatory pressure and compliance costs

Though regulations are getting tighter worldwide and recent fines have been substantial, it is clear financial institutions are struggling to fulfill their basic AML duties. The EU’s Anti-Money Laundering Authority (AMLA), which recently commenced operations in Frankfurt, is among the regulatory bodies tackling these challenges. Starting in 2028, AMLA will supervise high-risk financial organizations across the EU, ensuring uniform enforcement of rules and replacing the fragmented oversight previously carried out by individual member states.


Limitations of traditional AML systems

Traditional AML systems are largely built around static, rule-based models that trigger alerts when transactions exceed certain thresholds or match predefined scenarios. While these frameworks have served as the backbone of compliance operations for decades, they are increasingly showing their limitations.

One of the most persistent issues is the high rate of false positives—alerts that appear suspicious but ultimately prove benign. According to industry reports, including studies by Global Investigations Review and PricewaterhouseCoopers, as many as 95% of all alerts from current monitoring systems fall into this category. 

These erroneous alerts overwhelm compliance teams, divert resources from genuine threats, and can delay investigations into truly illicit activity. Rule-based systems struggle to keep up with the pace of innovation in financial crime, as criminal networks continuously adapt their laundering methods using complex structures or emerging technologies that fall outside the scope of hard-coded detection rules.


The emergence of AI and data analytics in AML

To counter these limitations, financial institutions are turning to artificial intelligence and data analytics. These technologies bring a more adaptive and intelligent approach to detection by analyzing vast volumes of data and recognizing complex or unusual behavior that may point to money laundering or fraud. Unlike static rule engines, machine learning models can learn and evolve over time, continuously refining their predictions based on real-world feedback.

AI can process historical transaction data to spot subtle red flags—such as irregular transaction patterns or deviations from a customer's typical financial behavior. Supervised learning algorithms can be trained to flag atypical fund transfers that would not necessarily trigger a traditional rule but are statistically unusual. Natural language processing tools can scan public and private sources—news reports, sanctions lists, legal filings, and even social media—for mentions of individuals or entities linked to financial crimes.

This shift toward AI-driven, adaptive detection methods enables institutions to move beyond reactive compliance into more proactive, intelligence-led risk management, ensuring quicker and more effective action against AML threats.

Customer Due Diligence (CDD) and Know Your Customer (KYC)

AI enhances CDD and KYC processes by automating identity verification and risk assessment. Machine learning algorithms can analyze customer data to detect inconsistencies or anomalies that may indicate fraudulent identities or high-risk profiles. 

Tools powered by natural language processing can analyze documents, social media, and open-source intelligence to uncover adverse media, sanctions, and ownership structures. This automation accelerates onboarding processes while maintaining compliance with regulatory standards.


Transaction monitoring

Traditional transaction monitoring systems often generate numerous false positives, burdening compliance teams with excessive alerts. AI-driven models can learn from historical transaction data to distinguish between legitimate and suspicious activities more effectively. By leveraging machine learning algorithms, banks can adapt to emerging threats, reduce false positives, and ensure compliance with evolving regulatory standards.

By identifying complex patterns and adapting to new laundering techniques, these models improve the precision of alerts. Machine learning algorithms can identify complex patterns and anomalies that may go unnoticed by traditional rule-based systems. By continuously learning from data, AI systems can improve detection accuracy and identify emerging money laundering techniques. AI doesn't replace existing rules; it enhances them. Many financial institutions now use a hybrid system: rules catch the known risks, while machine learning models look for emerging threats.

According to a study by McKinsey, machine learning can reduce false positives by over 50% while identifying 10-20% more alerts related to potential money laundering activities, freeing up resources for deeper investigations of genuine threats. HSBC, for example, which checks about 900 million transactions for signs of financial crime each month, found that AI implementation resulted in 60% fewer false positive cases.


Suspicious Activity Reporting (SAR)

Filing a SAR can take hours of manual effort. Some firms are now using AI to pre-fill SAR templates, suggest narrative language based on past filings, and extract relevant customer data automatically. While human oversight is still essential, the time savings are substantial.


Network analytics and graph-based approaches

Money laundering often involves complex networks of transactions across multiple entities and jurisdictions. Graph-based analytics enable the visualization and analysis of these networks, uncovering hidden relationships and transaction patterns that may indicate illicit activities. 

By representing entities as nodes and transactions as edges, graph models can identify central actors, detect community structures, and reveal indirect connections that traditional analysis might miss. AI-powered graph analysis tools are increasingly capable of finding suspicious clusters and relationships—like hidden beneficial ownership or circular transaction flows—that might never be detected through linear transaction reviews.


At a glance: What makes AI so effective?

• Pattern recognition: AI models excel at finding hidden relationships—something difficult or impossible to do manually. 

• Anomaly detection: When a transaction deviates from a customer's normal behavior in ways that don't trigger a rule, AI can flag it anyway. 

• Continuous learning: With the right feedback loops, machine learning systems can get smarter and more precise over time.


Challenges and considerations 

• Data quality and availability

The effectiveness of AI models depends on the quality and comprehensiveness of the data they are trained on. Incomplete, outdated, or biased data can impair model performance and lead to inaccurate predictions. Ensuring access to high-quality, representative datasets is crucial for the success of AI-driven AML systems. 

• Model interpretability

Regulatory compliance requires that financial institutions understand and explain the rationale behind AML decisions. Some AI models, particularly deep learning algorithms, operate as "black boxes," making it challenging to interpret their outputs. Developing explainable AI (XAI) techniques is essential to provide transparency and build trust in these systems. 

• Regulatory compliance

AI-driven AML solutions must align with existing regulatory frameworks, such as the Financial Action Task Force (FATF) recommendations and the Bank Secrecy Act (BSA). Ensuring that AI models adhere to legal standards and can be audited is critical for their adoption and effectiveness.

1 - Europe: EU AML Package and AMLA implementation

• Regulatory framework:
A new European Anti-Money Laundering Authority (AMLA) has been established in Frankfurt and formally began operations on July 1, 2025. With a staff of over 400 when fully operational, AMLA will focus on centralizing anti-money laundering efforts, coordinating national authorities and conducting cross-border investigations. The framework for these measures was laid out in the EU AML Package, a comprehensive set of legislative proposals adopted in 2021.

• Enforcement:
- Crypto-asset service providers are now required to collect and store information on the source and beneficiary of the funds for each transaction through the "travel rule".
- AMLA will have direct supervision over the "riskiest" obliged entities that work across borders, with the first selection process for “selected obliged entities” scheduled for July 2027, and direct supervision to begin in 2028.

• Regional fraud statistics:
In 2023, an investigation by the European Anti-Fraud Office found that over €1.2 billion had been affected by fraud and irregularities.

• Best practices:
- Implementing AI-based AML transaction monitoring for real-time and continuous monitoring
- Preparing for unified compliance platforms integrating all compliance functions
- Establishing enhanced beneficial ownership verification using trusted third-party data


2 - Asia-Pacific: Collaborative and technology-forward approach 

• Regional framework:
The Asia/Pacific Group on Money Laundering (APG) consists of 42 member jurisdictions, of which 12 are also FATF members: Australia, Canada, China, Hong Kong, India, Indonesia, Japan, Republic of Korea, Malaysia, New Zealand, Singapore, and the United States.

• Singapore leadership:
The Monetary Authority of Singapore launched its Collaborative Sharing of ML/TF Information & Cases (COSMIC) platform on April 1, 2024, co-created by six major international banks (DBS, OCBC, UOB, Standard Chartered, Citibank, and HSBC), facilitating information sharing between organizations with focus on misuse of legal persons, illicit trade finance, and proliferation financing.

• Regional fraud statistics:
According to Sumsub's 2023 Identity Fraud Report, Indonesia, Hong Kong, and Cambodia more than doubled their identity fraud percentages between 2021 and 2023.

• Best practices:
- Implementing collaborative information-sharing platforms like COSMIC
- Deploying AI and machine learning for transaction monitoring and pattern recognition
- Focusing on money mule detection through customer due diligence (CDD)
- Enhancing real-time payment fraud prevention systems

• Regional enforcement action:
Authorities in Hong Kong arrested 175 suspects linked to fraud and money laundering activities during a 16-day operation in 2024, involving fraud scams totaling around HK$780 million.


3 - North America: Risk-based and innovation-focused 

• Regulatory evolution:
On June 28, 2024, FinCEN announced a proposed rule that updates requirements for financial institutions' formal AML/CFT (Countering Financing of Terrorism) programs, introducing requirements that these programs be "effective, risk-based, and reasonably designed" and based on robust risk assessments.

• Key developments:
FinCEN will begin to provide access to beneficial ownership information (BOI) in phases to authorized government agencies and financial institutions, enhancing transparency and closing regulatory gaps in beneficial ownership reporting.

• Framework requirements:
Under the Proposed Rule, the risk assessment will become a fifth mandatory program element for all financial institutions, adding onto the four existing program obligations—internal controls, a qualified AML/CFT officer, training, and independent testing.

• Best practices:
- Implementing comprehensive risk assessment frameworks as foundation for AML programs
- Filing Currency Transaction Reports (CTRs) for transactions exceeding $10,000 in cash and Suspicious Activity Reports (SARs) for transactions of at least $5,000 that are flagged as suspicious
- Leveraging innovative approaches while maintaining regulatory compliance
- Focusing on beneficial ownership transparency and verification


4 - Middle East & Africa: Enforcement and compliance focus 

• Regional risk assessment:
Approximately one-third of jurisdictions in the Middle East and Africa are identified as high-risk, with 90% of the assessed territories listed as "money laundering jurisdictions" by the US. Assessments revealed significant differences across the region, indicating different institutional capacities to mitigate risks.

• UAE progress:
The FATF removed the UAE from the grey list in February 2024, acknowledging that the country had introduced substantial compliance measures and made progress in facilitating AML investigations and increasing prosecutions. Notable examples include the suspension of 50 companies for failing to register with the government's anti-AML system and issuing collective fines to financial institutions and real estate companies totaling more than AED 199 million (US$54 million).

• Best practices:
- Strengthening AML/TFS obligations with ongoing screening and reporting
- Implementing specialized federal prosecution divisions for economic crimes
- Enhancing due diligence procedures for high-risk customers
- Focusing on trade-based money laundering prevention


5 - Latin America: Drug trafficking and corruption focus 

• Regional challenges:
The majority of nations in Latin America exhibit moderate risks of money laundering and terrorist financing, with Haiti, Venezuela, and Suriname notably contributing to a decline in the regional average, and drug trafficking being the predominant source of money laundering risk.

• Key statistics:
Only three countries—Chile, Grenada, and Uruguay—escape inclusion in the US list of "major money laundering jurisdictions" in the context of drug trafficking, while there has also been an increase in risks associated with corruption and bribery. 

• Best practices:
- Implementing enhanced transaction monitoring for drug trafficking patterns
- Strengthening beneficial ownership verification for complex structures
- Focusing on corruption and bribery risk assessment
- Enhancing cross-border cooperation mechanisms

Looking ahead, the role of AI in AML is expected to advance in several key areas. Models that support continual learning will allow systems to adapt to new patterns without losing what they've already learned—critically important for spotting emerging laundering tactics.

Federated learning offers a way for institutions to collaborate and improve model performance without sharing sensitive data.

Meanwhile, as the use of cryptocurrencies grows, the integration of blockchain analytics will become essential for tracing illicit transactions in decentralized finance. The landscape of AML and fraud prevention continues to evolve rapidly, driven by advancing criminal methodologies and tightening regulatory requirements.

Financial institutions that successfully implement AI-driven, data-analytics-powered solutions while maintaining a holistic approach to financial crime prevention will be best positioned to protect their institutions, customers, and the broader financial system.

The future belongs to institutions that can effectively balance sophisticated AI capabilities with robust data governance, seamless integration across fraud and AML functions, and a proactive approach to emerging risks. As the stakes continue to rise, the question is not whether to invest in advanced AML and fraud prevention capabilities, but how quickly and effectively institutions can implement these critical safeguards while adapting to regional regulatory requirements and risk profiles.

• Global financial crime compliance costs: LexisNexis Risk Solutions, "True Cost of Financial Crime Compliance Study" - https://certpro.com/financial-crime-compliance/

• UNODC money laundering estimates: United Nations Office on Drugs and Crime, "Money-laundering Overview" - https://www.unodc.org/unodc/en/money-laundering/overview.html

• AI implementation trends: SymphonyAI, "What lies ahead for global financial crime prevention in 2025" - https://www.symphonyai.com/resources/blog/financial-services/financial-crime-prevention-2025/

• 2024 FTC Fraud Report: https://www.ftc.gov/news-events/news/press-releases/2025/03/new-ftc-data-show-big-jump-reported-losses-fraud-125-billion-2024

• 2023 FTC Fraud Report: https://www.ftc.gov/news-events/news/press-releases/2024/02/nationwide-fraud-losses-top-10-billion-2023-ftc-steps-efforts-protect-public

• 2022 FTC Fraud Report: https://www.ftc.gov/news-events/news/press-releases/2023/02/new-ftc-data-show-consumers-reported-losing-nearly-88-billion-scams-2022

• FinCEN Deepfake Alert (November 13, 2024): https://www.fincen.gov/news/news-releases/fincen-issues-alert-fraud-schemes-involving-deepfake-media-targeting-financial

• FinCEN Alert PDF: https://www.fincen.gov/sites/default/files/shared/FinCEN-Alert-DeepFakes-Alert508FINAL.pdf

• FinCEN record penalty: https://www.fincen.gov/news/news-releases/fincen-assesses-record-13-billion-penalty-against-td-bank

• AMLA establishment: https://en.wikipedia.org/wiki/Anti-Money_Laundering_Authority

• AMLA About page: https://www.amla.europa.eu/about-amla_en

• AMLA Frankfurt location: https://www.europarl.europa.eu/news/en/press-room/20240219IPR17818/frankfurt-will-be-the-home-of-the-eu-anti-money-laundering-authority

• AMLA 2028 supervision timeline: https://eucrim.eu/news/tasks-powers-and-structures-of-amla/

• Thomson Reuters historical data (2017): https://www.reuters.com/article/us-banks-regulator-fines/u-s-eu-fines-on-banks-misconduct-to-top-400-billion-by-2020-report-idUSKCN1C210B/

• Global Investigations Review (2020): https://blackdotsolutions.com/blog/aml-alerts/

• PricewaterhouseCoopers False Positives Report: https://www.datavisor.com/blog/guest-post-end-the-false-positive-alerts-plague-in-anti-money-laundering-aml-systems/

• McKinsey on AML challenges: https://www.mckinsey.com/capabilities/risk-and-resilience/our-insights/the-investigator-centered-approach-to-financial-crime-doing-what-matters

• FATF official website: https://www.fatf-gafi.org/

• Academic research on transaction monitoring: https://www.sciencedirect.com/science/article/pii/S0167739X24002607

• Lucinity. "Understanding False Positives in Transaction Monitoring: What Causes

• Them and How AI Can Reduce Operational Waste." November 6, 2024. https://lucinity.com/blog/understanding-false-positives-in-transaction-monitoring-what-causes-them-and-how-can-ai-can-reduce-operational-waste

• Tookitaki. "Smart Surveillance: How AI is Revolutionizing Transaction Monitoring." https://www.tookitaki.com/compliance-hub/ai-transaction-monitoring-real-time-compliance

• Financial Crime Academy. "AI Revolution In AML: Optimizing Transaction Monitoring For Compliance." May 13, 2025. https://financialcrimeacademy.org/aml-transaction-monitoring-using-ai/

• HSBC. "Harnessing the Power of AI to Fight Financial Crime." June 10, 2024. https://www.hsbc.com/news-and-views/views/hsbc-views/harnessing-the-power-of-ai-to-fight-financial-crime

• Lucinity. "Exploring the Evolution of AML Transaction Monitoring." April 30, 2024. https://lucinity.com/blog/exploring-the-evolution-of-aml-transaction-monitoring

• Financial Crime Academy. "Combatting Financial Crimes: AML Transaction Monitoring in Banking." May 22, 2025. https://financialcrimeacademy.org/aml-transaction-monitoring-in-banking/