With the help of machine learning models, underwriting in lending has created compelling applications across use cases. Several techniques are being used by businesses to develop, manage, explain, audit and manage their models. They are streamlining operations, automating processes and making data-driven decisions.
Designing a governance framework around these Underwriting AI models is important, as:
- Failure or errors can have a serious impact on the profitability and quality of the book
- There are directly applicable regulations to use AI, like the AI Act (US, EU), GDPR (EU), RBI Digital Guidelines (India), etc., with extensive guidance outlining the expectations of regulators
- Identifying and mitigating bias in underwriting has become crucial
- Auditability is important for risk management and regulatory compliance
- The productivity of underwriters can be improved if the models are transparent
In this workshop, we walk through a case study on how ML Observability tools can be used to design a framework for achieving a scalable and safe AI environment for the business.
You will learn:
- Overview of Lending Club Data and loan default prediction model
- Overview of ML observability, applications and best practices
- How to use AryaXAI and design the AI Governance on Lending club data
- Next steps and future work