Model Governance Full Disclosure – Using Resilient AI/ML Systems with Confidence

The value of sophisticated model risk management extends well beyond the satisfaction of regulatory regimes. But how can we ensure that their frameworks are capturing this value thoroughly? Join us to find out the answers! Our speakers deep dive into using resilient Artificial Intelligence (AI) and Machine Learning systems with confidence. Governance for confidence: How to deal with Black Box Models How can you use AI/ML models without knowing what is inside them? Many model explainability methods make good guesses but may not meet the confidence levels of managers who are used to knowing exact variables and their contributions. This session explores how managers can design governance structures and processes to provide them with the confidence to use models with unknown variables. Building resilient AI systems that you can trust In an age of automation and digitalization, the use of Artificial Intelligence and Machine Learning (AI and ML) is now mainstream in our society. It is also delivering tangible benefits in Risk Management: AI and ML are able to improve the accuracy of risk estimation models, automate repetitive processes, and accelerate risk-based decision making. However, ‘brittle’ AI and ML models are costly to develop and validate and do not transfer well to real-world applications. How can firms build more trust and confidence in AI and ML by improving model resiliency, explainability and putting in place the right controls? The session highlights practical cases where AI and ML deliver value to the Risk function and discusses global industry practices on how robustness assessments can be incorporated in continuous monitoring frameworks.