ACAMS Netherlands Chapter: How to Build Ethical and Trusted AI/ML Models to combat FinCrime

The ACAMS Netherlands Chapter is hosting a webinar about building an ethical and trusted AI/ML (Artificial Intelligence/Machine Learning) models to combat financial crime. This webinar will focus on the guidelines from the European Commission and the Bank for International Settlements and actionable steps for financial institutions to implement these guidelines. Recently, national regulators have also expressed concerns about this topic and a regulatory guidance can be expected in near future.

Increasingly, financial institutions are using AI/ML models (to provide input) for decision-making. Next to applications in credit risk, application of AI/ML models in Financial Crime processes has intensified significantly. In these processes, AI/ML models have an increasing influence and share in customer onboarding, due diligence, adverse media screening, and alert investigation. However, the current system holds some significant challenges around designing, implementing and interpreting these AI/ML models.

This webinar explores the challenges and considerations in developing a trusted and responsible AI model. The webinar will consist of two sessions of 20 minutes and 10 mins of discussion and intends to discuss the following major areas:

1. Policy and governance for trusted AI models: How to develop trust among different lines of business, customers, and regulators about the AI/ML models and make these models scalable for business processes?
2. Diversity and fairness in AI/ML models: Are institutions maintaining the right mix of academic, gender, racial diversity in model developer's teams in order to not implicate/continue the societal biases?
3. Explain-ability and scalability: Has institution implemented appropriate governance and explain-ability measures for the AI/ML Black box to reduce the biases, reduce overfitting and identify whether the inferences make a cognitive sense and are causally relevant?