Commonwealth Scientific and Industrial Research Organisation

CSIRO adopts the Australian Government’s Policy for the Responsible Use of AI in Government, applying the Digital Transformation Agency’s framework to ensure our AI practices remain safe, ethical and transparent. (Template language)

CSIRO has been using AI for many years in the form of machine learning, to support scientific discovery. CSIRO is committed to responsibly expanding its use of AI to improve operational efficiency and support future scientific developments. Planned initiatives include applying AI tools to streamline administrative and data heavy workflows, reduce manual processing times, strengthen data analysis for enhanced insights, and develop digital interfaces using natural language processing to make internal information more accessible.

CSIRO uses AI to deliver the science Australians need for the nation they want – productive, sustainable, healthy and secure. We also use AI to improve our organisational efficiency. Our approach is guided by Australia’s AI Ethics Principles, including:

• Human, societal and environmental wellbeing (Shared with 1 other agency)
• Privacy protection and security (Shared with 1 other agency)

All AI use at CSIRO, whether in research or enterprise operations, is subject to human oversight. Outputs are reviewed by qualified personnel to ensure accuracy and appropriateness, commonly referred to as ‘human-in-the-loop’ (HITL) governance.

CSIRO's Chief Digital Officer is the designated accountable official for AI use at CSIRO. (Template language)

CSIRO applies AI across both research and enterprise domains. Current usage includes:

Pattern | Description
---|---
Decision making and administrative action | Supports human decision-making through recommendations and assessments.
Analytics for insights | Uses AI and related methods (including machine learning, deep learning, and agent-based approaches) to analyse scientific datasets and identify patterns.
Workplace productivity | Assists with tasks such as content creation, summarisation, workflow optimisation, data analysis, and secretariat support.

Pattern | Description
---|---
Scientific | Enhances research through data processing, simulation, prediction, and monitoring.
Corporate and enabling | Supports HR, finance, communications, IT, and all staff through automation and optimisation.

Public Impact and Safeguards

CSIRO does not currently use AI in a way that allows direct interaction with the public or significantly impacts individuals without human oversight. Existing AI systems are limited to internal functions that support data processing, analysis, and workflow efficiency. Despite this low level of exposure, CSIRO applies safeguards to prevent any potential negative impacts, using the DTA Policy to consider risks to privacy, fairness, and inclusivity—even when systems only operate within internal environments.

We are committed to reviewing this position as technology and use cases evolve.

All research undertaken with human subjects is governed by CSIRO's Human Research Ethics Committee and conducted in compliance with the principles and requirements specified in the National Statement on Ethical Conduct in Human Research 2023

Oversight of data and AI governance is provided by the Data and AI Governance Committee, with operational activities undertaken by the Data and AI Governance team.

To support AI governance, CSIRO is implementing:

• A Responsible Use of AI Procedure, aligned with internal frameworks such as the Code of Conduct procedure, Privacy Policy, and Responsible Use of ICT Procedure.
• A training module on Responsible use of AI.

Staff are encouraged to raise any concerns or seek guidance on AI use through established internal channels.

CSIRO recognises that AI is an evolving field. We will introduce regular internal reviews of all AI systems to evaluate accuracy and intended impacts. These reviews will inform continuous improvement.

We will continue to monitor, evaluate, and refine our approach to AI use, ensuring alignment with whole-of-government initiatives, community expectations and relevant legislation. This statement will be updated as required to reflect changes in AI usage and governance.

Publication date: 6 November 2025

Last Update date: 6 November 2025