EVIDENT AI BANKING INDEXEVIDENT AI INSURANCE INDEXINSIGHTSNEWSLETTEREVENTSABOUT USJOIN USSIGN IN
EIA-rai-hero

How are banks adapting governance structures to ensure responsible AI deployment?

Tuesday, March 11, 2025

15:00-16:00 GMT / 11:00-12:00 EDT

Virtual Roundtable

Back to Events
RET-headshot_1

Raquel Ettrick Thompson

Global Head of AI Governance & Control, BNY

Raquel oversees AI/ML governance, risk management, and compliance across the enterprise at BNY Mellon. She leads efforts to ensure the trustworthiness and regulatory alignment of AI-driven solutions while managing controls for Enterprise Payments functions.

Paul Dongha

Dr. Paul Dongha

Head of Responsible AI & AI Strategy, NatWest

Paul leads the development and implementation of ethical AI practices and strategies across the Group. Paul's work ensures AI technologies are designed and used in line with regulatory standards and ethical guidelines, mitigating risks related to bias, transparency, and accountability.

Alexandra Mousavizadeh

Co-founder & Co-CEO, Evident

Alexandra has spent the last 25 years ranking and quantifying complex societal and political forces. She started her career at Moody’s, then Morgan Stanley, later became CEO of ARC Ratings and then Director of the Prosperity Index. Most recently, Alexandra was a Founding Partner at Tortoise Media, where she ran Tortoise Intelligence, the Index and data business. Here, she was the architect of the groundbreaking Global AI Index, the first benchmark to track the strength of national AI ecosystems. She holds a degree in economics, mathematics and game theory from the University of Copenhagen.

Professional headshot of a smiling man with a beard and short dark hair, wearing a grey plaid blazer over a blue collared shirt.

Colin Gilbert

VP of Intelligence, Evident

Colin leads Evident's Intelligence Team. focused on developing new content offerings and expanding our customer-facing diagnostics. Colin has spent the last 15 years working in Research & Advisory services, specializing in using quantitative benchmarking to measure an organization's digital aptitude vs. performance. Previously, he served as a Managing Vice President at L2 Inc. (acquired by Gartner in 2017), where he worked on the Digital IQ Index® as it expanded to cover nearly 1,500 leading brands across 15x B2C industries.

Watch the Roundtable

Summary

Scaling Generative AI use cases across banking comes with the underlying challenge to address the complexity and “black-box” nature of models, whilst maintaining the robust oversight and controls standards set over years of dealing with technology risk.

To prevent risk management and governance becoming a blocker to innovation, leading banks are proactively adapting their People, Policies and Process to ensure they deploy AI models and systems responsibly.

Our recent RAI Roundtable offered a unique opportunity to learn about the practical measures taken by subject matter experts from some of the world’s leading banks to ensure RAI supports, not stifles, innovation.

This event focused on the updated 2024 Evident AI RAI Report.

Key Discussion Topics

  • The role of the RAI leader: why they’re needed and what they do
  • Execution of RAI principles and mapping them to discrete controls
  • How to “future proof” controls, especially in tandem with dedicated research into AI explainability (XAI)
  • Key processes to help overcome the bottleneck challenge of validating AI models and scaling Generative AI use cases

Key Slides

Data visualization showing bank performance in AI transparency for 2023-2024. A table lists the top 15 performing banks in the Transparency Pillar of the Evident AI Index, with JPMorgan Chase ranked first in both overall index and transparency. A bar chart titled 'Change in Average Index Score, by Pillar (2023 vs. 2024, n=50 Banks)' shows Transparency with a +27.2% increase, Leadership with +23.4%, Overall Score with +8.0%, and both Innovation and Talent with +2.7% gains.

Responsible AI (RAI) maturity framework showing progression across six levels (Emerging to Pioneering) for People, Policy, and Process dimensions. The framework details specific characteristics at each stage of RAI implementation, with levels 4 and 5 highlighted as the current position for leading banks.

Quote from Dr. Paul Dongha, Head of Responsible AI & AI Strategy at Natwest: “So as much as we like to talk about AI’s huge potential and how it’s going to transform people’s lives and give customers much better experiences, I think we have to give equal importance to the risks and how we can mitigate them, because we can mitigate them. It’s perfectly doable, but it needs mobilization across the organization – inclusive of people, process, technology, and policies…”

Data visualization presenting the growth and composition of Responsible AI (RAI) talent in banks, which increased 41% year-over-year. The left stacked bar chart shows RAI talent volume by region (France, Europe, APAC, UK, Canada, USA) growing from approximately 175 units in 2023 to 250 units in 2024. The right bar chart highlights an increase in the percentage of banks with RAI leaders, rising from 30% in 2023 to 66% in 2024.

Quote by Raquel Ettrick Thompson, Global Head of AI Governance & Control at BNY: 'In terms of leadership, there has to be someone that can bridge that middle – someone that is able to translate to the business and describe why the capabilities, the internal activities, the governance, and the working groups need to be cross-functional. They have to understand the expected outcome, managing the business value and benefit of AI alongside core company principles.'

Table detailing AI risk and governance committee structures in banks, with a title stating 'PEOPLE: 29 of 50 banks have established AI Risk or Governance Committees. Profiles of cross-team stakeholders that serve on AI Committees & Working Groups.' The table lists five teams: 1. Responsible AI Risk & Compliance Team, 2. AI Governance & Policy Team, 3. AI Model Oversight & Assurance Team, 4. AI Ethics & Responsible Innovation Team, and 5. Technology & AI Security Team. For each team, specific roles (e.g., AI Risk Lead, AI Ethicist, Model Risk Manager) and their corresponding responsibilities (e.g., conduct risk assessments, ensure AI models align with regulations, develop AI risk mitigation strategies, oversee AI model validation, establish AI incident response framework) are outlined.

Quote by Raquel Ettrick Thompson, Global Head of AI Governance & Control at BNY: 'A big part of what I see as Responsible AI, is about making sure that you have the right stakeholders at the table to consistently evaluate, not only the industry expectations, the regulatory expectations, the legal expectations, but how it aligns to your company’s values and direction.'

Presentation slide titled 'PROCESS: Toward (Automated) AI Assurance Platforms - Evidence of iterative, incremental progress across leading banks'. A table shows AI assurance initiatives: CommBank is implementing automated model checks (implemented); CIBC is embedding AI governance in MLOps (implemented); BNY is enhancing monitoring dashboards for anomaly detection (iterative); and NatWest is building a company-wide AI assurance platform (pilot). A sidebar defines AI Assurance as the process organizations use to measure and demonstrate system trustworthiness to government, regulators, or the market, referencing the UK Department of Science, Innovation, and Technology (updated February 2024).

Quote by Dr. Paul Dongha, Head of Responsible AI & AI Strategy at Natwest: 'A lot of a lot of people think: ‘Oh, my God, all this extra regulation and risk management – it’s slowing us down…’ I don’t believe that for one minute. I actually think, certainly for large organizations, it gives you certainty and clarity over what you can proceed with, and then you can proceed at pace. I think that level of certainty, knowing that the investment [in RAI] is a worthwhile investment that helps overcome regulatory hurdles down the line … I think that’s really, really important.' Below the quote is a circular photo of Dr. Paul Dongha.