With the new government-wide policy on safe, responsible use of artificial intelligence (AI) now in effect, the Digital Transformation Agency (DTA) has provided further details as to how it will continue to support Australian Public Service (APS) agencies through skills training, standards and assurance frameworks.

The DTA indicated that it would continue to take into account the rapid changes of AI technologies and update its policies to remain aligned with any advancements and changing community expectations.

The agency also revealed its four-pronged approach to support APS agencies in their pathway to safe and responsible AI use, including AI fundamentals training, developing an AI assurance framework, key takeaways from the Microsoft Copilot trial and implementing new AI technical standards.

“It is important for APS agencies to provide staff with the knowledge and support needed to determine if general-use AI tools, including generative AI, are right for their work,” a statement said.

“The policy strongly recommends that agencies implement AI fundamentals training for all their staff and, by early October, the DTA will publish an AI fundamentals training module.

“This training offers a clear, consistent understanding of how AI works and what to consider when making responsible choices. It will be available for agencies to integrate into their learning platforms, as well as on APSLearn.

“We will also refresh the ‘Guidance for using generative AI’, a simpler rename of the Interim guidance for agency use of public generative AI tools, to be clearer, easier to access and a great companion to the fundamentals training.”

The DTA confirmed it will trial an AI assurance framework for the Australian Government to standardise risk management processes around AI use cases, off the back of the national, state and territory agreement made in June to the National framework for the assurance of artificial intelligence in government.

The agency also said it would share key takeaways from its Copilot for Microsoft 365 trial with the public.

“General-use AI capabilities, such as generative AI, may impact many agencies, their staff and ways of working. It creates new opportunities as well as new risks,” the statement said.

“The government has faced this change proactively, including through its cross-APS trial of Copilot for Microsoft 365. These findings will be relevant to both agencies considering using generative AI capabilities and vendors looking to offer these capabilities to government.

“To encourage responsible innovation, DTA is developing AI technical standards, bringing together the APS’ leading AI and machine learning expertise through a cross-government working group, with work continuing through to mid 2025.

“The resulting standards will be published under an open licence. This continues our commitment to transparency and encourages both consistency and interoperability between governments, academia and industry as we all contribute to safe and responsible AI in Australia.”

This also comes as the Australian National Audit Office (ANAO) announced it had extended the due date for contributions to its audit of the Australian Taxation Office’s (ATO’s) adoption of artificial intelligence until 15 September.