What is AI model risk management (MRM)?
Financial institutions use AI models to predict and measure qualitative information. Supervisory guidance such as SR-11-7 (Link resides outside IBM) requires that models be tested and managed for risk. IBM Watson® OpenScale™, unlike many MRM solutions, automates model testing throughout the model lifecycle and syncs results with systems of record. It can help detect and correct AI model bias and drift and explain AI outcomes. This simplifies validation, especially for multiple models. Watson OpenScale is included in IBM Cloud Pak™ for Data and available on IBM Cloud™. It works with models and data running virtually anywhere.
Review key screens in AI model risk management
This Watson OpenScale dashboard shows results for fairness, quality and drift tests, two of which did not clear custom thresholds. Users can drill down for further details.
Users can configure and customize a variety of tests to fit their business needs.
This screen shows how users can compare model test results for a variety of criteria side-by-side. This can speed up and formalize the development of more effective models.
To enhance compliance, Watson OpenScale can automatically generate a PDF fact sheet summarizing details about the model, related data, and test results.
Explore model risk management
5 ways that Watson OpenScale simplifies AI model risk management
Customize tests and thresholds with IBM Watson OpenScale to enhance model compliance.
451 Research recognizes Watson OpenScale for innovation
Read how IBM Watson OpenScale can help you implement AI at scale, while building trust in AI via monitoring, fairness and explainability.
How the game has changed in AI model risk management
Read a review of the new Watson OpenScale MRM capabilities.
Inside the MRM workflow
Explore how Watson OpenScale works with IBM OpenPages® Model Risk Governance
Get a quick guided click-through of how to automate tests and automatically document results.
Webinar: AI model risk management challenges and solutions
Explore model validation challenges with two model validation leaders at RBC and learn how tests can now be automated and streamlined.
Read about the new factors generating model risk
In an era of big data and AI, new kinds of validation tests are needed.
Where is MRM headed?
Streamlining MRM throughout the model lifecycle
Create challenger models faster. Compare results and automatically document them.
Listen to where MRM is headed
AI experts Kelly Combs from KPMG and Rohan Vaidyanathan from IBM discuss industry trends in how to make AI more trustworthy.
Learn from MRM documentation
Explore details on the range of capabilities for automating AI model tests.