Singapore has released a software toolkit aimed at helping financial institutions ensure they are using artificial intelligence (AI) responsibly. Five whitepapers also have been issued to guide them on assessing their deployment based on predefined principles.
The Monetary Authority of Singapore (MAS) said the documents detailed methodologies for incorporating the FEAT principles--of Fairness, Ethics, Accountability, and Transparency--into the use of AI within the financial services sector.
The whitepapers were developed by the Veritas consortium, which is part of Singapore's national AI strategy and comprises 27 industry players that include Amazon Web Services, Bank of China, Bank of Singapore, Google Cloud, Goldman Sachs, OCBC Bank, and Unionbank of the Philippines.
According to MAS, the whitepapers provide a FEAT checklist to guide financial institutions in their AI and data analytics software development lifecycles as well as an enhanced fairness assessment methodology to define the objectives of their AI and data analytics systems and identify potential bias.
There also is a methodology for assessing ethics and accountability, which offers a framework to help financial institutions carry out quantifiable measurement of ethical practices, and another for assessing transparency so these organisations can determine how much internal and external transparency is needed to explain and interpret predictions generated by machine learning models.
The Veritas consortium also developed the software toolkit to automate the fairness metrics assessment and facilitate visualisation of the assessment interface. Available on GitHub, the open source toolkit allows for plugins to enable integration with the financial institution's IT systems.
MAS' chief fintech officer Sopnendu Mohanty said in the statement released Friday: "The new open source software, assessment methodologies, and enhanced guidance will further improve the technical capabilities of financial institutions in developing responsible AI for the financial sector."
Some members of the Veritas consortium also applied the methodologies to various functions within their organisation, including customer marketing, insurance fraud detection, and credit risk scoring.
The group next would develop additional use cases and conduct pilots with selected financial institutions within the consortium to further integrate the methodologies with their existing governance framework.
MAS added that it was working with Infocomm Media Development Authority and Personal Data Protection Commission (PDPC) to include the tolkit in the PDPC's Trustworthy AI testing framework.