MAS sets guidelines for responsible AI and data analytics use

It aims to promote fairness, ethics, accountability and transparency in using and managing data.

The Monetary Authority of Singapore (MAS) released a set of principles to promote fairness, ethics, accountability and transparency (FEAT) when using artificial intelligence (AI) and data analytics in finance, an announcement revealed.

Known as the ‘FEAT principles’, it aims to promote responsible use of AI and data analytics (AIDA) to firms offering financial products and services, as well as to strengthen internal governance around data use and management.

MAS noted that it worked closely with a group of senior industry partners in developing the principles through a Feat Committee which incorporated the views and feedback from financial institutions, industry associations, fintech firms and technology providers. The committee is co-chaired by MAS chief data officer David Hardoon and PrimePartners co-founder and advisor Hsieh Fu Hua.

“As the financial industry harnesses the potential of AI and data analytics, we need to be cognisant of using these technologies in a responsible and ethical manner,” Hardoon said.

Under ‘fairness’, MAS noted that data and models used for AIDA-driven decisions must be regularly reviewed and validated for accuracy and relevance to minimise unintentional bias, whilst under ‘ethics’, AIDA-based decisions and use must be aligned with the firm’s ethical standards and hold the same standards as human driven decisions.

Accountability is divided into internal and external under the guidelines, and MAS highlighted that firms using AIDA must inform their respective boards on its use, and that data subjects are provided with the necessary channels to enquire about AIDA-driven decisions that affect them.

Use of AIDA must be proactively disclosed to data subjects as part of general communication, including what the data is used for and its consequences to data subjects, to increase public confidence under ‘transparency’.

The full principles document is available on the MAS website.