Almost a year to the day since the Singapore government released its first edition of the Model AI Governance Framework, a second edition has been announced at the 2020 World Economic Forum (WEF), aimed at improving consumer trust on the use and governance of artificial intelligence (AI).
The updated version [PDF] of the Model AI Governance Framework includes considerations such as robustness and reproducibility, to make it more relevant and usable.
For instance, the section on customer relationship management has been expanded to include considerations on interactions and communications with a broader network of stakeholders, and the need to consider the purpose and context when interacting with the various stakeholders.
Other updates include the addition of industry examples by companies such as Facebook and Mastercard to illustrate how organisations have implemented sections of AI governance practices.
Clarifications around algorithm audits were also added, with the second edition noting an algorithm audit should only be conducted "if it is necessary to discover the actual operations of algorithms comprised in models, and only at the request of a regulator (as part of a forensic investigation)".
Similarly, the second edition clarified the "human-over-the-loop" approach by explaining the human's supervisory role in AI-augmented decision-making and that organisations can consider other factors, such as the nature and reversibility of harm, and operational feasibility to determine the level of human involvement in an organisation's decision-making process around AI.
"The objective really is to translate these ethical principles -- that it must be human-centric and responsible -- and translate that into practical guidelines so that companies who want to adopt AI technologies have a kind of guidebook on how they can go about doing this whilst maintaining the trust of their clients and their customers," Singapore Minister for Communications and Information S Iswaran said.
According to Iswaran, some 15 organisations have already adopted the framework because they share a combined objective to make AI practices "human centric" and "transparent".
"It must ensure, therefore, that it engenders trust so that people can then use the technologies, or benefit from the technologies, and not be burdened by concerns. This is the context for much of the work that we are doing in Singapore, including the model AI Governance Framework," he said.
"The key is, we want to continue to build on this momentum. The organisations are international, and what it shows is that they see value in the framework."
The launch of the second edition is one of three new initiatives the Singapore government launched during WEF. The other two initiatives include an implementation and self-assessment guide for organisations (ISAGO) and a compendium of use cases.
Developed with input from more than 60 organisations, including Google, Microsoft, Salesforce, Mastercard, and Visa, ISAGO [PDF], the Singapore government said, is intended to help organisations assess the alignment of their AI governance practices with the Model AI Governance Framework by providing industry examples and practices to help organisations implement it.
Meanwhile, the compendium of use cases [PDF] shows how local and international organisations across different sectors and sizes have implemented or aligned their AI governance practices with the framework and benefited from it. Some of these companies include Visa and HSBC.
"Ultimately, organisations benefit from seeing how others have done it. So when you have these use cases, they become a very valuable tool; helping them to contextualise it to their own environment," Iswaran said.
The latest edition of the framework builds on Singapore's unveiling of its national AI strategy last November as part of plans to drive the adoption of AI to create economic value, enhance citizen lives, and arm its workforce with necessary skills.