Home Artificial Intelligence Singapore authority publishes ‘methodologies’ for responsible AI

Singapore authority publishes ‘methodologies’ for responsible AI

Singapore: the methodologies have been compiled by MAS and more than 20 private companies working collectively as a consortium called ‘Veritas’ | Credit: Pexels; Pixabay

Assessment methodologies to steer and encourage responsible use of artificial intelligence (AI) by financial institutions such as commercial banks have been published by the Monetary Authority of Singapore (MAS).

The methodologies have been compiled by MAS and more than 20 private companies working collectively as a consortium called ‘Veritas’. Their publication marks the culmination of the second phase of a collaboration that kicked off in November 2019 after the authority’s release of principles to promote fairness, ethics, accountability and transparency (‘FEAT’) in the use of AI and data analytics (‘AIDA’) across Singapore’s financial sector in 2018.

The newly released papers include a checklist across FEAT principles for financial institutions to adopt during AI and data analytics software development; and a ‘Transparency Assessment Methodology’ that seeks to help financial institutions to determine whether and how much internal and external transparency is needed to ‘explain and interpret’ the predictions of machine-learning models.

The public-private consortium has also just released the first version of an open-source software toolkit, which similarly aims to drive financial institutions’ adoption and adherence to FEAT methodologies and principles. The software enables the automation of metrics assessment and visualisation, with plug-ins integrating with financial institutions’ IT systems.

Veritas is set against the backdrop of regulators worldwide striving to tackle the opportunities and challenges presented by AI and data. Examples include the European Union’s proposal for a regulation laying down harmonised AI rules (announced in April 2021) and, in the UK, the Bank of England and Financial Conduct Authority’s establishment of an ‘Artificial Intelligence Public Private Forum’ to discuss the use and impact of AI in financial services.

Global Government Fintech lists Veritas consortium members at the end of this article

MAS’s eyes on ‘fostering innovation’

MAS’s chief fintech officer, Sopnendu Mohanty, said the Veritas initiative would “further improve the technical capabilities of financial institutions in developing responsible AI for the financial sector”. He said the initiative “continues to deliver tangible outcomes that demonstrate collaborative public-private partnership to drive trust in the adoption of AI technology, enhance confidence and foster innovation in Singapore’s fintech ecosystem”.

In January last year, MAS published two documents to mark the end of the first phase of the Veritas project: a ‘FEAT Fairness Principles Assessment Methodology’ and ‘FEAT Fairness Principles Assessment Case Studies’. The focus during this first phase was credit-risk scoring and customer marketing for the banking industry.

Four teams within the consortium developed the methodologies, open-source software toolkit and also use-case studies in the banking and insurance sectors.

Specifically, Swiss Re and Accenture refined the ‘Fairness Assessment Methodology’ and applied it to insurance predictive underwriting; United Overseas Bank (UOB), AXA and Accenture developed an ‘Ethics and Accountability Assessment Methodology’ and applied it to customer marketing and insurance fraud detection; Standard Chartered Bank, HSBC and Truera developed the Transparency Assessment Methodology and applied it to credit-risk scoring and customer marketing; and Accenture and Bank of China developed the open-source software toolkit with support from BNY Mellon, HSBC, Standard Chartered and UOB. 

Looking ahead, the consortium plans to develop additional use cases and run pilots to integrate the methodologies with members’ governance frameworks. It will also work on a second version of the toolkit to provide ‘full FEAT assessment functionality’.

MAS is also collaborating with the Singapore’s Infocomm Media Development Authority and Personal Data Protection Commission (PDPC) to include the toolkit in the PDPC’s Trustworthy AI testing framework.

Veritas initiative’s global aspect

The Veritas initiative also has a globally-oriented dimension.

Specifically, the most recent Singapore FinTech Festival saw MAS announce the winners of what was billed as the inaugural ‘Global Veritas Challenge’. This was a competition to highlight and accelerate the development of fintech solutions that ‘validate artificial intelligence and data analytics solutions against the FEAT principles, to strengthen trust and promote greater adoption of AI solutions in the financial sector’.

The three winners, selected from ten finalists during a ‘demo day’ (20 October 2021), addressed challenges in validating the fairness of artificial intelligence (AI) solutions.

The challenge focused on four use cases: product marketing; risk, compliance and fraud monitoring; loan origination and know-your-customer (KYC); and credit scoring and profiling. There were 74 submissions.

Announcing the winners at the festival in November, Mohanty said the organisers looked forward to seeing further innovative solutions “in future editions” of the competition.

VERITAS CONSORTIUM: who’s in it?

1) Accenture 2) Amazon Web Services 3) AXA 4) Bank of China 5) Bank of Singapore 6) BAFT (Bankers Association for Finance & Trade) 7) BNP Paribas 8) BNY Mellon 9) CFA Society Singapore 10) Citi Singapore 11) DBS Bank 12) Google Cloud 13) Goldman Sachs 14) HSBC 15) IBM 16) Industrial and Commercial Bank of China Limited 17) Microsoft 18) Monetary Authority of Singapore 19) MUFG Bank Ltd 20) OCBC Bank 21) Ping An of China 22) Prudential Assurance Company Singapore 23) Standard Chartered Bank 24) Swiss Re 25) Truera 26) Unionbank of the Philippines 27) United Overseas Bank Limited

FURTHER READING

‘East Asian governments surge in AI readiness – see global rankings in full’ an article in our sister title Global Government Forum published on 26 January 2022 on the ‘2021 Government AI Readiness Index’ (from Oxford Insights)

‘A closer look at Singapore’s AI governance framework: insights for other governments’ – an article in our sister title Global Government Forum published on 6 May 2021