1.
Introduction & Research Background
Topic: Enhancing AML Risk Management in Banking: A Qualitative and
Quantitative Study on AI-Driven Financial Crime Detection
Anti-Money Laundering (AML) compliance is a critical operational risk for banks,
driven by increasing regulatory scrutiny and the complexity of financial crimes.
Traditional rule-based AML models often generate high false positives, resulting
in inefficiencies and regulatory challenges. Artificial Intelligence (AI) and
advanced risk analytics offer potential solutions to improve financial crime
detection, but concerns around regulatory compliance, model interpretability,
and governance remain barriers to widespread adoption.
This dissertation aims to explore the challenges of AML compliance through a
hybrid research approach, combining qualitative insights from industry
professionals with quantitative testing of AI-driven models.
2. Research Objectives
Primary (Qualitative) Objectives:
Investigate the operational challenges in AML compliance and risk
governance within financial institutions.
Conduct expert interviews with AML professionals on the adoption of AI in
risk frameworks.
Analyze regulatory perspectives on the integration of AI-driven AML
models.
Secondary (Quantitative) Objectives:
Compare the performance of AI-based AML systems with traditional rule-
based models.
Apply machine learning models (e.g., Random Forest, XGBoost) to
evaluate whether AI can effectively reduce false positives in fraud
detection.
Conduct statistical hypothesis testing (e.g., Chi-square test, ROC-AUC) to
validate AI’s effectiveness.
3. Research Questions
1. What are the key operational challenges in AML compliance and risk
governance for banks?
2. How do financial institutions perceive AI adoption in AML frameworks, and
what regulatory concerns exist?
3. Can AI-based AML models reduce false positives without compromising
compliance?
4. What role do regulatory frameworks play in shaping AI adoption for
financial crime detection?
4. Research Methodology
Phase 1: Qualitative Approach
Data Collection: Conduct interviews with AML professionals, compliance
officers, and financial risk managers.
Analysis: Thematic analysis of expert insights and regulatory documents.
Output: Develop a robust understanding of the challenges and
opportunities for AI in AML risk management.
Phase 2: Quantitative Approach (If Needed)
Data Source: Utilize datasets like SAML-D or synthetic AML transaction
data.
Modeling Techniques: Implement machine learning models to assess
AI’s performance in fraud detection.
Evaluation Metrics: Accuracy, precision, recall, F1-score, and ROC-AUC.
Hypothesis Testing: Use statistical tests to validate improvements in
operational risk management.
5. Initial Literature Review
Regulatory Frameworks: Review of FATF AML guidelines, EU AML
directives, and relevant banking compliance standards.
AI in AML: Analysis of existing AI-based risk management models and
their performance compared to traditional systems.
Operational Risk Management: Identification of key risk management
practices in the banking sector related to AML compliance.
6. Research Timeline & Key Milestones
Milestone Description Deadline
Discuss initial proposal and gather
First Meeting February 2025
feedback
Refinement of Research Questions Finalize based on supervisor feedback March 2025
Dissertation Proposal Submission 28th March
Submit 2,000-3,000 word proposal
(20%) 2025
Qualitative Research Completion Conduct interviews and analyze data April-May 2025
Quantitative Analysis (If
Build and test AI models May-June 2025
Required)
7. Proposed Meeting Agenda
1. Introduce Myself: Academic and professional background in financial
risk management.
2. Present Dissertation Proposal: Outline key research objectives and
methodology.
3. Seek Guidance: Clarify specific areas such as research focus, interview
protocols, and potential quantitative analysis.
4. Confirm Expectations: Discuss frequency of meetings, milestones, and
preferred communication methods.
8. Specific Questions for Dr. Smatralova
Do you recommend any specific academic frameworks for analyzing AML
compliance challenges?
Are there particular regulatory documents or case studies I should
prioritize during my qualitative research?
Should I prepare an ethics application for conducting interviews with AML
professionals?
Would you advise focusing more on the qualitative phase initially and
expanding to quantitative only if needed?
9. Next Steps
Incorporate feedback from our meeting into the proposal.
Start conducting a literature review and preparing interview questions.
Develop a preliminary list of AML professionals for qualitative interviews.
10. Conclusion
This research aims to bridge the gap between regulatory compliance and AI-
driven AML risk management through a structured hybrid approach. I am
eager to gain your insights and guidance to refine this study further and ensure
it meets both academic and industry standards.
I look forward to our meeting and thank you in advance for your support and
supervision.
Best regards, Arun Joy Perumana MSc Financial Risk Management | Student
ID: 24348167 [Your Contact Information]
