Singapore has announced the launch of the world’s first governance framework for agentic artificial intelligence (AI). The Model AI Governance Framework (MGF) for Agentic AI was announced at the World Economic Forum (WEF) last week. While it might have been overshadowed by certain other topics and speakers at the forum, it represents a significant moment in AI deployment and governance.
The new framework builds on the city state’s existing MGF for AI, which was first introduced in 2020, but focuses on agentic AI. Unlike generative AI and other models that perform limited tasks, agentic AI is designed to perceive, reason, plan and act autonomously to carry out multi-step processes and achieve complex goals. As AI agents are increasingly embedded into organisations’ operations, they are likely to be given access to sensitive data and the ability to make changes to their environment, such as updating customer databases or making payments. While this could prove to be even more transformative than generative AI, some warn that it could also pose significant risks.
Framework aims to manage risks and maintain human accountability
Singapore’s Infocomm Media Development Authority (IMDA), which drew up the framework, stated: “The increased capability and autonomy of agents also create challenges for effective human accountability, such as greater automation bias, or the tendency to over-trust an automated system that has performed reliably in the past.”
The MGF for Agentic AI emphasises that human oversight should be maintained throughout the lifecycle of a system or series of tasks, with humans ultimately remaining accountable for all actions and outcomes. The recommended steps to achieve this goal include defining limits on the extent of an agentic AI system’s independence and the access it has to data and tools. Checkpoints where human approval is required should be defined and established at relevant points of the system. The framework also advises that continuous monitoring and baseline testing should be carried out throughout the lifecycle of the agentic AI.
The new MGF is targeted at all organisations looking to adopt and deploy agentic AI, whether this involves developing the agents in-house or using third-party models. In a statement accompanying the launch, the IMDA said that it was intended to provide organisations with guidance on the technical and non-technical measures they need to deploy the technology responsibly.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
Destination AI
Destination AI from TD SYNNEX is our comprehensive end-to-end programme designed to help you and your customers get ahead of the AI curve.
