Search

Classification system for AI use

Version 1.1

The following classification system represents how AI is commonly used in government and the domains where they are applied. Agencies must list both the usage patterns and domains which apply to their use of AI in their transparency statements.

Agencies should link to this page when listing their usage patterns and domains and may expand on their use of AI to provide more details to the public.

Usage patterns

Decision making and administrative action

Used to either:

• support decision making or the taking of administrative action by guiding, assessing or making a recommendation to a human decision maker
• make decisions or take administrative action without human intervention. 

Note: not all automated decision making may be considered AI, noting the definition under the policy.

Examples include eligibility decision systems for government services or making recommendations based on submitted applications.

Analytics for insights

Used to identify, produce or understand insights within structured or unstructured materials via comprehensive data analysis, predictive modelling and/or reporting tools.

Examples include risk detection systems and forecast modelling.

Workplace productivity

Automates routine tasks, manage workflows, and facilitate communication.

Examples include virtual assistants, automated scheduling, content creation and basic secretariat support.

Image processing

Processes images to automatically identify patterns and objects, such as faces, for official purposes.

Examples include facial recognition for border control and identifying pests in cargo to aid biosecurity.

Domains

Service delivery

Enhances efficiency or accuracy of government services, including payment services, by providing tailored and responsive services to the public.

This may include in direct interaction with the public, such as chat-bots, enhanced customer self-service and multilingual capabilities, or support staff or systems which deliver services.

Compliance and fraud detection

Identifies patterns or anomalies in data to detect fraudulent activities and ensure compliance with laws and regulations.

Law enforcement, intelligence and security

Supports law enforcement and intelligence agencies by analysing data from various sources to predict and prevent crimes, and by aiding in intelligence gathering.

Policy and legal

Analyses policies and legal documents to provide advice and assurance on their impact and supports policy development that is consistent with existing laws.

Scientific

Leveraged in scientific endeavours to process complex datasets, simulate experiments, predict outcomes and enhance monitoring functions.

Corporate and enabling

Supports corporate functions, including HR, finance, media and communications, and IT, by automating processes, optimising resource allocation and improving operational efficiency.