Technical

Agentic AI

AI systems capable of autonomous action, planning, and executing multi-step tasks with minimal human intervention.

How the laws define this

Regulatory instruments universally define agentic AI as systems designed to act autonomously, make decisions, and execute actions to achieve objectives without constant human oversight or direction. Where these texts primary diverge is in whether they frame the concept around general technical capabilities or specific public-sector applications. Broader frameworks—such as the Government response to the AI regulation white paper (AI regulation: government response), the Selangor AI White Paper, and the Updated Model AI Governance Framework for Agentic AI—focus on fundamental operational mechanics like breaking down tasks, tool selection, adapting to feedback, and addressing regulatory safety considerations. In contrast, implementation guidance like the Agentic AI for Government Services Framework and Agentic AI for Government Services grounds the definition directly in public service delivery, emphasizing the automated management of government workflows, coordination with authorities, and the direct execution of administrative transactions on behalf of citizens.

Synthesised from the 5 statutory definitions below. Each one is quoted in full, with its source.

How jurisdictions differ

What is distinctive in each definition, binding law in force first.

Jurisdiction	Instrument	Status	What is distinctive
Singapore	Updated Model AI Governance Framework for Agentic AI	In Force	Adds breaking down tasks into subtasks, selecting tools, executing actions, adapting to real-time feedback, and interacting with other agents and systems.
United Arab Emirates	Agentic AI for Government Services Framework	In Force	Focuses on government operations, emphasizing proactive actions, managing complex workflows, supporting decision-making, completing transactions, and optimizing services.
United Kingdom	A pro-innovation approach to AI regulation: government response	Adopted	Emphasizes extended sequences of steps or interactions and pursuing objectives or sub-goals without continual human direction, highlighting relevance to safety assessment.
Malaysia	Selangor AI White Paper	Adopted	Limits to systems specifically engineered for predefined objectives, stressing advanced decision-making and execution capabilities without broader task decomposition or interactions.
United Arab Emirates	Agentic AI for Government Services	In Force	Centers on public services, highlighting understanding user needs, coordinating with authorities, and completing transactions or tasks on behalf of citizens.

What it triggers

In Singapore, the Updated Model AI Governance Framework sets out guidance for organizations deploying agentic AI to manage systemic risks and maintain human oversight.source
In the UAE, the Agentic AI for Government Services Framework guides integration of agentic AI into government services to boost efficiency and oversight.
In the United Kingdom, the government response guides sectoral regulators on applying cross-sectoral AI principles to agentic AI capabilities as relevant to safety assessment.source
In Malaysia's Selangor, the AI White Paper outlines a strategic framework for development and adoption of agentic AI.source

Definitions (5)

AI systems characterised by the capacity to autonomously plan, make decisions, and act over extended sequences of steps or interactions, potentially pursuing objectives or sub-goals without continual human direction; the Response flags agentic capabilities as a distinct class relevant to safety assessment and regulatory consideration.

Government response to the AI regulation white paper (AI regulation: government response)Definition 1 of 5

Agentic AI refers to AI systems specifically designed to operate autonomously and proactively. These systems are engineered to achieve predefined objectives without constant human intervention, demonstrating advanced decision-making and execution capabilities.

Selangor AI White PaperDefinition 2 of 5

Artificial intelligence systems capable of autonomous planning, reasoning, and action, which can break down tasks into subtasks, select tools, execute actions, and adapt to real-time feedback, often interacting with other agents and systems to complete tasks on behalf of humans.

Updated Model AI Governance Framework for Agentic AIDefinition 3 of 5

Agentic AI refers to sophisticated artificial intelligence systems that not only process information but also proactively perform actions, manage complex operational workflows, and support decision-making with minimal human intervention. These systems are designed to operate as independent agents within government operations, taking steps to complete transactions and optimize services based on predefined objectives and data analysis.

Agentic AI for Government Services FrameworkDefinition 4 of 5

Agentic AI refers to advanced artificial intelligence systems designed to operate autonomously, manage complex operations, and execute a series of independent actions without requiring constant human oversight. These systems are characterized by their ability to understand user needs, coordinate with relevant authorities, and complete transactions or tasks on behalf of citizens, streamlining public services.

Agentic AI for Government ServicesDefinition 5 of 5

Not to be confused with

Human Oversight

Human oversight refers to mechanisms for control and intervention, while agentic AI definitions describe capabilities for minimal intervention or autonomous operation.

Meaningful Human Control

Meaningful human control is a principle requiring human understanding and authority, contrasting with agentic AI's emphasis on autonomy with minimal or no continual direction.

Questions and answers

What guidance applies to deploying agentic AI in Singapore?

The Updated Model AI Governance Framework guides organizations to manage systemic risks and maintain human oversight when deploying agentic AI agents. It sets out technical measures for these non-binding recommendations.

How does the UK approach regulation of agentic AI?

The government response adopts a pro-innovation approach, guiding sectoral regulators to apply cross-sectoral principles to agentic AI systems flagged for safety assessment due to their autonomous capabilities.

What is the focus of agentic AI in UAE government services?

Agentic AI systems are designed to autonomously manage operations, understand user needs, coordinate with authorities, and complete tasks on behalf of citizens to streamline public services.