25 February 2022

On 4 February 2022, MAS announced the release of five white papers detailing assessment methodologies for the Fairness, Ethics, Accountability and Transparency (“FEAT”) principles to guide the responsible use of AI by financial institutions (“FI”). The white papers were published by the Veritas Consortium (“Consortium”), comprising 27 industry players. The white papers provide the following:

  • a comprehensive FEAT checklist for FIs to adopt during their Artificial Intelligence and Data Analytics (“AIDA”) software development lifecycles;
  • an enhanced Fairness Assessment Methodology to enable FIs to define their AIDA system’s fairness objectives, identify personal attributes of individuals and any unintentional bias;
  • a new Ethics and Accountability Assessment Methodology, which provides a framework for FIs to carry out quantifiable measurement of ethical practices, in addition to the qualitative practices currently adopted;
  • a new Transparency Assessment Methodology which helps FIs determine whether and how much internal/external transparency is needed to explain and interpret the predictions of machine learning models.

The Consortium has also developed an open-source software toolkit to accelerate FIs’ adoption of the FEAT methodologies and principles. The toolkit enables the automation of the fairness metrics assessment and allows for visualisation of the interface for fairness assessment and for the plug-ins to integrate with the IT systems of FIs.

In the next phase, the Consortium will develop additional use cases and run pilots with selected FI members to integrate the methodologies with members’ existing governance framework. MAS is also collaborating with the Info-communication Media Development Authority and the Personal Data Protection Commission (“PDPC”) to include the toolkit in the PDPC’s Trustworthy AI testing framework.

Reference materials

The following materials are available on the MAS website www.mas.gov.sg.