The Central Bank of Nigeria's (CBN) new Baseline Standards for Automated Anti-Money Laundering (AML) Solutions, while lauded as among the world's best, introduce a complex set of risks that financial institutions must navigate. The effectiveness of these standards hinges on their implementation, with the CBN emphasizing demonstrable effectiveness over mere feature-based compliance.

These new standards, designed to enhance AML efforts, also present ten significant risks that require careful attention from banks. Understanding these risks and implementing appropriate governance is crucial for genuine compliance.

Understanding the Risks in AML Compliance

The CBN has been explicit: compliance must be effective, not just a tick-box exercise. This directive underscores the importance of how banks integrate and utilize these new AML solutions.

One of the primary risks identified is Algorithmic Bias. AI models used for customer risk scoring rely on attributes like geography, occupation, income, transaction channel, and customer segment. If a model is trained predominantly on data from urban, formally employed, high-income customers, it may systematically score customers outside this profile as higher risk, simply because their behaviour is statistically unfamiliar.

This bias has significant implications in Nigeria's diverse financial landscape, serving informal traders, agricultural producers, and diaspora remittance recipients. Their transaction patterns can differ vastly from those of a typical Lagos salary earner, making biased algorithms a legal and ethical concern.

Another critical risk is Model Drift, where the performance of a model degrades over time as real-world data evolves. Banks must have mechanisms to monitor and retrain models to ensure their continued accuracy and relevance.

Explainability Failure poses a challenge, as complex AI models can become