Artificial Intelligence Assurance Framework

The Northern Territory (NT) Government  Artificial Intelligence (AI) Assurance Framework (November 2025) outlines guidelines for the responsible use of AI in government to enhance service delivery while managing associated risks.

The framework aims to improve public services through responsible AI use.

This AI Assurance framework provides ethics principles, a self-assurance assessment and a quick risk assessment to help agencies:

  • authorise the use of AI and provide guidance on best practice and risks
  • meet the expectations of Territorians in developing safe and responsible AI
  • respect the unique challenges, risks and opportunities presented by AI
  • encourage innovative approaches to service delivery and decision-making.

Purpose

The framework is to be used when NT Government agencies or government owned corporations are designing, building or using AI-enabled systems.

It focusses on ensuring that AI is used to enhance service delivery for Territorians in ways that are safe, responsible and ethical to maintain public trust.

To ensure the framework does not stifle innovation, it applies stronger controls for medium and high-risk activities, while simplifying measures for low-risk activities.

What is AI

AI refers to technology that simulates human intelligence, enabling computers and machines to perform tasks that typically require human cognition.

It is useful to view AI not as a single technology, but as a broad spectrum of tools with varying levels of complexity and impact.

At one end are everyday AI tools like spam filters, predictive text, and streaming recommendations. They are low-risk, human-controlled, and designed to make life easier.

At the other end are high-risk AI uses like predictive policing, facial recognition, and autonomous systems making critical decisions. If not carefully governed, these controversial uses of AI raise serious concerns about fairness, privacy, and accountability.

AI ethics principles for responsible use

The framework has ethics principles to guide AI use.

Community benefit AI must serve the public interest by delivering demonstrable, positive outcomes for the community.
Safety AI must be used safely and responsibly and should reliably operate as intended.
Fairness Use of AI systems must be proactively and continuously managed to identify, mitigate and include safeguards against bias, preventing the risk of data bias, or creating or perpetuating unjust outcomes.
Privacy and security AI systems must incorporate the highest levels of security and assurance to safeguard all forms of sensitive information, including personal data, commercial-in-confidence material and Cabinet-related information.
Transparency Decisions and outputs from AI systems must be understandable and explainable to those they affect, in proportion to the level of risk. Clear mechanisms must be available for individuals to question and challenge AI-assisted outcomes.
Accountability Public sector organisations and public sector staff retain ultimate responsibility for decisions using AI. This requires meaningful human oversight and clear lines of accountability for the outcomes of AI systems.

Assurance process for AI projects

A structured assurance process is outlined for AI projects.

  • The agency chief executive is accountable for all ICT projects, including AI.
  • A quick risk assessment determines the initial risk level of AI projects.
  • Low-risk projects can proceed with agency-level oversight, while medium and high-risk projects require additional assurance steps supported by a specialised AI advisory service within NT Government.

Definitions

Artificial Intelligence (AI)The NT Government defines artificial intelligence (AI) as intelligent technology, programs and the use of advanced computing algorithms that can augment decision making by identifying meaningful patterns in data and creating new content. AI in this context should aim to help government agencies cut costs, free up labour hours for more critical tasks, and deliver better, more targeted services.
AI projectAn ICT project that develops, acquires, implements or adopts AI, including when the project team is acquiring and implementing a larger product suite and an AI product is built into that product and required or intended to be used.
Personal data or personal information

As defined in the Information Act 2002, means NT Government information that discloses a person’s identity or from which a person’s identity is reasonably ascertainable.

Sensitive information

As defined in the Information Act, means:

  • personal information about:
    • racial or ethnic origin
    • political opinions
    • membership of a political association
    • religious beliefs or affiliations
    • philosophical beliefs
    • membership of a professional or trade association
    • membership of a trade union
    • sexual preferences or practices
    • a criminal record
  • health information.