Article ID: b49f918e810767cdfce80e24a7cec9b77bd7ab4b753e155f1ef22f40e54f2b10
Source ID: regulatory:risk.net
Published At: -
Extraction Method: trafilatura
URL: https://www.risk.net/resource/7963435/mrm-how-banks-are-scaling-models-in-the-age-of-ai?cta=true
Body Text
MRM: How banks are scaling models in the age of AI As artificial intelligence and regulatory change drive rapid growth in model inventories across banks, model risk management is entering a new phase. Institutions must strengthen governance, improve efficiency and maintain control over increasingly complex model ecosystems – without slowing innovation. In this white paper, produced by Risk.net in association with Moody’s, senior industry practitioners explain how banks are adapting their model risk management (MRM) frameworks to meet rising expectations around inventory management, validation, AI oversight and lifecycle governance. Drawing on a Risk.net survey of 79 model development, validation and risk professionals, alongside interviews with senior experts from Standard Chartered, Citigroup, CaixaBank, EastWest Bank, Nordea and Moody’s, the report explores how firms are responding to the surge in models across the enterprise. Key themes include: - Why AI, machine learning and regulation are accelerating model growth across banking - How banks are modernising model inventories and governance frameworks - The biggest bottlenecks in development, validation and deployment - How institutions are integrating AI/machine learning models into existing MRM frameworks - The role of centralised platforms, automation and workflow tools in scaling MRM - What the future model risk function might look like in an AI-enabled banking environment. As model volumes rise and governance expectations broaden, banks face a clear challenge: scaling controls quickly enough to remain compliant while preserving speed, innovation and competitive advantage. Download the paper to learn how leading banks are evolving MRM for the AI era. Download the whitepaper Register for free access to hundreds of resources.
Metadata (JSON)
{
"score": 4.333333333333333
}