Veritas Consortium AI Fairness Testing Methodology

Applies To

Overview

The Veritas Consortium was established by the Monetary Authority of Singapore (MAS) in 2019 to translate its Fairness, Ethics, Accountability and Transparency (FEAT) Principles for financial institutions into concrete, testable assessment methodologies. The consortium has included major banks, insurers, and fintech firms operating in Singapore. The Veritas methodology provides financial institutions with a practical toolkit for assessing AI-driven use cases against each of the four FEAT dimensions. Phase 1, published on 6 January 2021, built a fairness methodology for two banking use cases: credit risk scoring and customer marketing. Phase 2, published on 4 February 2022, covered all four FEAT principles and added insurance underwriting and fraud detection, with an open-source toolkit. Toolkit 2.0 and a Phase 3 white paper followed on 26 June 2023. The Veritas framework puts FEAT into practice through numerical fairness measures, documentation requirements, and governance process checks tailored to the data and model types common in financial services. Using it is voluntary: MAS presents Veritas as an open-source toolkit that helps financial institutions assess their AI against the FEAT principles.

Key Requirements

• •Veritas helps financial institutions assess AI use cases against the four FEAT principles: Fairness, Ethics, Accountability, and Transparency.
• •Fairness assessments involve choosing quantitative metrics suited to the use case and documenting why.
• •Published case studies cover credit risk scoring and customer marketing in banking, and predictive underwriting and fraud detection in insurance.
• •Assessments also look at governance, such as human oversight and model lifecycle management.
• •The open-source Veritas Toolkit supports these assessments, and using it is voluntary.
• •Reassessing systems as models or deployment contexts change is good practice.

What Your Organization Must Do

• →Map your AI use cases in financial services against the four FEAT principles: fairness, ethics, accountability and transparency.
• →Download the open source Veritas Toolkit and test it on one credit or marketing model first.
• →Choose fairness measures that fit each use case and record in writing why you picked them.
• →Document human oversight and model lifecycle controls for each AI system you assess.
• →Reassess systems when models change or when you deploy them in a new context.
• →Tell your board that Veritas is voluntary guidance, not a binding Singapore legal requirement.

Playbook Guidance

Step-by-step implementation guidance for compliance teams.