Agenda
Day 1
4.6.2025
6.4.2025
-
Registration and Breakfast
-
Opening Address from the Chair
-
ADAPT MODEL RISK FRAMEWORKS FOR AI AND GEN AI INTEGRATION STRENGTHENING GOVERNANCE OVERSIGHT AND ETHICAL DECISION-MAKING
-
Keynote
Reshape model risk frameworks and methodologies considering the advancements of AI, ML, and Gen AI
- Examine how the introduction of AI, ML, and Gen AI models is reshaping traditional model risk frameworks and assessment methodologies
- Discuss the evolving definition of model risk, including challenges related to data integrity, privacy, explainability, and transparency in AI-driven models
- Explore strategies for adapting validation and evaluation processes to effectively assess the accuracy and reliability of AI model
Lee Medoff
Head of Model Governance
Fifth Third Bank
Against
Lee Medoff
Head of Model Governance
Fifth Third Bank
Panelists
Lee Medoff
Head of Model Governance
Fifth Third Bank
-
Balance the adoption of AI and traditional models to strengthen governance and ensure effective oversight
- Explore the best practices for integrating AI and non-AI models within model risk frameworks to ensure consistency and oversight
- Analyze the challenges in maintaining and improving traditional models amidst the rise of AI and ML technologies
- Assess the effectiveness of leveraging traditional models as benchmarks to assess the efficacy of emerging AI models
-
Networking and Refreshment Break
-
Overcome the challenge of defining and categorizing Generative AI outputs in model frameworks
- Review the distinction between models and tools, addressing challenges in defining AI-driven models and ensuring appropriate regulatory compliance
- Discuss the challenges of categorizing Gen AI models within traditional model risk frameworks and their implications for governance
- Examine strategies to enhance transparency in Gen AI model outputs, addressing opacity and data capture limitations
- Explore approaches for integrating Gen AI models into existing risk management frameworks while maintaining regulatory compliance
-
Case Study
Explore techniques to identify and mitigate bias in AI models to ensure fair and ethical decision-making
- Identify strategies to harmonize traditional risk management models with AI approaches for enhanced accuracy and reliability
- Address the challenges involved in effectively integrating AI models with traditional risk management frameworks
- Discover techniques for mitigating bias in AI models to ensure fair and ethical decision-making in risk management
-
Lunch
-
ENSURE REGULATORY COMPLIANCE IS MAINTAINED DESPITE ADVANCED AI MODELS INTEGRATION AND UNCERTAINTIES
-
Explore supervisory perspectives on model risk management to ensure compliance is maintained
David Palmer
Senior Supervisory Financial Analyst
Federal Reserve Board
-
Networking and Refreshment Break
-
NETWORKING AND GOUP INNOVATION EXCHANGES
-
Roundtable
Compare criteria and effective approaches for classifying and validating the institution’s model inventory
- What are the most accurate strategies used to distinguish models and tools within the institution's model inventory?
- What are the steps taken to manage the resource implications of classifying algorithms as models?
- What approaches are effective for upscaling validation functions and summarizing model risk reports?
Roundtable
Evaluate challenges and opportunities in leveraging LLMs for risk management and compliance
- What unique challenges do LLMs present regarding data integrity, transparency, and model explainability, and how can these risks be mitigated?
- In what ways can LLMs enhance model documentation, validation, and automation in regulatory reporting?
- What best practices and lessons can be learned from early adopters integrating LLMs into existing model risk frameworks and governance structures?
Ken King
Risk Director of AI
Citizens
Against
Ken King
Risk Director of AI
Citizens
Panelists
Ken King
Risk Director of AI
Citizens
Roundtable
Discuss strategies to improve real-time monitoring for enhanced risk management
- How can AI be used to continuously monitor model performance and detect anomalies in real time?
- In what ways can AI help identify emerging risks and model degradation, enabling proactive risk management and early intervention?
- How can AI-driven monitoring be integrated with traditional governance and compliance frameworks to ensure alignment and transparency?
-
Closing Comments from the Chair and End of Day 1
Day 2
5.6.2025
6.5.2025
-
Registration and Breakfast
-
Opening Address from the Chair
-
OPTIMIZE MODEL INVENTORY TO ENHANCE GOVERNANCE AND INTEGRATE AI WITH TRADITIONAL MODELS
-
Keynote
Improve model inventory workflows for streamlined risk management and effective oversight
- Examine how expanded departmental access to model inventories affects workflow complexity and risk oversight
- Investigate best practices for integrating model inventories with version control and monitoring tools to streamline processes
- Discuss strategies to improve collaboration between model risk teams and developers for a more cohesive risk management approach
Ken King
Risk Director of AI
Citizens
Against
Ken King
Risk Director of AI
Citizens
Panelists
Ken King
Risk Director of AI
Citizens
-
Strengthen model inventory governance and resource allocation to enhance integration and overcome complexities
- Assess the impact of increasing model inventory complexity on resource allocation and governance frameworks
- Examine the inefficiencies of static model inventories and explore integration with monitoring platforms and version control tools
- Review strategies for dynamic model inventory management to enhance oversight and cross-department collaboration
-
Networking and Refreshment Break
-
Panel Discussion
Compare strategies for integrating AI and traditional models into a unified model risk management framework
- How can organizations successfully align AI and non-AI models within current governance structures to enhance model risk management?
- What are the primary challenges in managing hybrid model inventories, and how can these be mitigated to ensure consistency and accuracy across all models?
- What best practices can be adopted to ensure regulatory compliance for both AI and traditional models within the broader risk management framework?
Alexey Smurov
Head of Line of Business Model Risk Management and Administration
PNC
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
Aashish Prakash
Head Model Risk Management
American Express
Against
Alexey Smurov
Head of Line of Business Model Risk Management and Administration
PNC
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
Aashish Prakash
Head Model Risk Management
American Express
Panelists
Alexey Smurov
Head of Line of Business Model Risk Management and Administration
PNC
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
Aashish Prakash
Head Model Risk Management
American Express
-
ENSURE REGULATORY COMPLIANCE IS MAINTAINED DESPITE ADVANCED AI MODELS INTEGRATION AND UNCERTAINTIES
-
Analyze the implication of different regulatory requirements for financial institutions operating globally and its impact on model risk management standards
- Examine the implications of the EU AI Act for financial institutions in Europe and contrast its impact on global model risk management with regulatory frameworks in other regions
- Discuss how transparency and legal integrity in AI and ML model governance are maintained across varying regulatory landscapes
- Evaluate best practices for aligning model documentation, validation, and oversight with evolving regulations
-
Lunch
-
Panel Discussion
Ensure ongoing alignment and compliance with SR11-7 regulations despite evolving model risk management frameworks
- What steps can be taken to adapt model risk management frameworks to align with SR11/7 in light of evolving technological advancements?
- What best practices should be followed to establish a robust framework that ensures continuous compliance with SR11/7, with a focus on documentation, validation, and governance?
- What strategies can be adopted to effectively monitor model performance and maintain SR11/7 compliance throughout the model lifecycle?
- What are the common challenges in meeting SR11/7 requirements, and what solutions can mitigate regulatory and operational risks in model risk management?
Julia Litvinova
Head of Global Advisors Model Risk and Corporate Model Validation, Managing Director
State Street
Han Cheng
Director, Model Risk Oversight
Charles Schwab
Brad Curell
Executive Director – Model Risk
Ally
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
Against
Julia Litvinova
Head of Global Advisors Model Risk and Corporate Model Validation, Managing Director
State Street
Han Cheng
Director, Model Risk Oversight
Charles Schwab
Brad Curell
Executive Director – Model Risk
Ally
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
Panelists
Julia Litvinova
Head of Global Advisors Model Risk and Corporate Model Validation, Managing Director
State Street
Han Cheng
Director, Model Risk Oversight
Charles Schwab
Brad Curell
Executive Director – Model Risk
Ally
Tina Tian
Senior Vice President, Head of Model Risk Management
Synchrony
-
ENHANCE MODEL VALIDATION AND OVERSIGHT BY ADDRESSING TRANSPARENCY AND CROSS-DEPARTMENT COLLABORATION CHALLENGES
-
Advance AI model validation overcoming transparency and testing challenges
- Analyze the unique challenges of validating AI and ML models, including transparency, bias mitigation, and explainability
- Review best practices for testing AI models and setting validation standards to ensure governance is maintained
- Examine the role of AI and Gen AI in automating documentation and validation processes to enhance efficiency
-
Networking and Refreshment Break
-
Case Study
Enhance collaboration across departments to strengthen AI and ML model risk oversight
- Discuss the importance of cross-department collaboration in managing AI and ML model risks, particularly between data science and risk teams
- Explore strategies for talent development and knowledge-sharing to refine AI model governance and oversight
- Assess the impact of cloud-based ML development on model access, validation, and regulatory compliance
Dan Saunders
Head of Model Risk Management
USAA
Against
Dan Saunders
Head of Model Risk Management
USAA
Panelists
Dan Saunders
Head of Model Risk Management
USAA
-
Panel Discussion
Discuss the complexities of assessing qualitative AI models and the need for standardized validation frameworks
- What are the challenges in assessing qualitative AI models, particularly those providing non-quantitative, predictive, or interpretive outputs?
- Why is there a need for standardized validation frameworks to ensure consistency, transparency, and regulatory compliance in evaluating qualitative AI models?
- What are the best practices for establishing criteria to assess the accuracy, fairness, and explainability of qualitative AI models across different industries?
- How can cross-functional teams collaborate effectively in the development and validation of qualitative AI models to ensure comprehensive risk management and alignment with governance frameworks?
-
Closing Comments from the Chair and End of Day 2
Post Conference Workshop
6.6.2025
6.6.2025
-
Registration and Breakfast
-
Workshop
Enhance model criticality tiering for effective risk management and ongoing compliance
- Identify key parameters that define a model’s criticality, including its impact on business decisions, regulatory compliance, and financial stability
- Examine frameworks and methodologies for assigning risk ratings to models, considering factors such as complexity, data quality, and usage frequency
- Understand the importance of ongoing monitoring and periodic reassessment to ensure critical models remain accurate and relevant over time
- Review case studies and best practices for successfully implementing model criticality assessments and risk tiering processes
-
Networking and Refreshment Break
-
Workshop
Build a resilient model risk ecosystem by strengthening governance and collaboration across AI, ML, and traditional models
- Explore the key components of a model risk ecosystem and how they interact to ensure effective risk management across the organization
- Discuss the role of cross-functional collaboration between departments like risk, data science, compliance, and IT in building a robust model risk ecosystem
- Investigate strategies for integrating AI, ML, and traditional models within the ecosystem to ensure consistent oversight and governance across different model types
- Assess the challenges of maintaining compliance and transparency in a dynamic model risk ecosystem, focusing on evolving regulatory requirements and industry best practices
-
End of Workshops