Australian Department of Finance Releases New Guidance on the Responsible Use of Generative AI
The guidance applies to tools including Microsoft 365 Copilot, Microsoft 365 Copilot Chat, and public AI platforms such as ChatGPT, Claude and Gemini.
Recognising the growing role of AI in improving productivity and service delivery, the guidance emphasises that AI should support, not replace, human judgement. Staff are required to maintain oversight of AI-generated content, verify outputs for accuracy, and remain accountable for decisions and actions.
The guidance also highlights the importance of privacy, security and information governance. It establishes clear rules around what information can be entered into different AI tools, with stricter controls applying to personal, protected and sensitive information. Public AI tools are restricted to unclassified information, while approved enterprise tools such as Microsoft 365 Copilot may be used with broader categories of information under specific conditions.
For records and information management professionals, the guidance reinforces several key principles, including the need for transparency, human accountability, appropriate recordkeeping, and compliance with privacy and security requirements. Staff are also directed to follow guidance from the National Archives of Australia when managing records created using AI.
The Department has also established an AI Governance Committee to oversee the adoption and use of AI, ensuring compliance with Australian Government policies, ethical principles and legislative requirements.
As AI adoption continues to expand across government, the guidance provides a practical example of how organisations can balance innovation with governance, helping ensure AI is used safely, responsibly and in a way that maintains trust in information and decision-making processes.