MAS unveils 3 winners of global challenge to promote AI adoption in financial sector

A TOOLKIT that helps financial institutions improve fairness of their marketing practices, an open-source governance framework for building reliable and fair machine learning models, and a solution to mitigate credit scoring biases were the 3 "responsible artificial intelligence (AI)" solutions that won the Monetary Authority of Singapore's (MAS) Global Veritas Challenge, which sought to promote the adoption of AI solutions in the financial sector.

These 3 winning projects - backed respectively by Visa, fraud detection solutions provider Cylynx, and US-headquartered TruEra and Demyst - were selected from a pool of 10 finalists. Each winner received a cash prize of S$50,000.

This year's challenge centred on 4 use cases, including product marketing; risk, compliance and fraud monitoring; loan origination and know-your-customer; and credit scoring and profiling.

Validating the fairness of AI and data analytics solutions in these areas ensures that these solutions do not result in particular groups of customers being systematically disadvantaged, the MAS said.

Visa's winning entry provides financial institutions with a set of reusable components for screening and mitigating unwanted biases in marketing models.

Cylynx's solution, which comes under the risk, compliance and fraud monitoring theme, helps business teams flag individuals or groups that may be systematically disadvantaged. It can also set thresholds to avoid unintended biases.

The final winning solution, a collaboration between US-headquartered firms TruEra and Demyst, looks at mitigating the effects of bias in credit scoring by using external data to build a machine learning-based credit scoring model. It helps users identify root causes leading to bias and accuracy issues and conduct targeted interventions.

Sopnendu Mohanty, chief fintech officer of the MAS, noted the importance of "responsible AI" in fostering trust among stakeholders.

"Responsible AI is essential to fostering trust among relevant stakeholders including users of AI solutions and recipients of AI-driven decisions. The solutions allow financial institutions to harness the tremendous benefits of AI in their operations while guarding against risks," he said.