Statement of Intent on AI in the Australian Public Service
Artificial Intelligence (AI) technologies are rapidly evolving and already transforming the way we do business. These tools offer unprecedented opportunities to deliver better services for Australians and drive public service efficiency. The challenge for us is to leverage these technologies for the benefit of all Australians while protecting them from harm. Adoption of AI technologies by government should always be for the benefit of people – not for its own sake.
We want to use AI to:
- make people’s lives better
- improve government service delivery, efficiency, and productivity
- ensure the benefits of this technology are shared equitably, creating a higher standard of living for all Australians.
This approach aligns with the Government’s ambition for AI in Australia, to:
- capture the opportunity of AI
- spread the benefits widely, and
- keep Australians safe.
The Australian Public Service (APS) is uniquely positioned to support Government’s AI ambitions by leading from the front. By procuring, developing and using safe AI tools, the APS can support Australia’s AI market, build the social license for AI use, ensure widespread access and inclusion and demonstrate practical management of AI risks.
As per the Data and Digital Government Strategy, the Australian Government is committed to a modern public service that puts people and business at the centre of it's data and digital transformation. The work of the AI Plan for the Australian Public Service is a step towards that vision.
The Government is committed to demonstrating the highest standards of safe and ethical use of AI in the APS. Public confidence and trust are essential for the Government to embrace AI’s potential. The community rightly expects the Government to use new technologies in a manner that is safe, transparent, and responsible. The Government will always put people first and ensure AI adoption doesn’t exacerbate inequality, bias, or unfair treatment.
Building on strong foundations
There are a range of existing legal and policy protections in place that safeguard the rights of Australians:
- The APS Values and Code of Conduct already provides base level expectations of public servants across all their work, regardless of what technology they use to help deliver it.
- The Protective Security Policy Framework prescribes what Australian Government entities must do to protect their people, information and resources.
- The Privacy Act 1988 promotes and protects the privacy of individuals and regulates how Australian Government agencies handle personal information.
- There are also governance mechanisms such as audit processes, the Commonwealth Ombudsman, and the Australian Human Rights Commission, all of which provide channels for issues to be proactively raised.
AI specific measures
However, there are also a range of additional AI-specific obligations and guidance to support this intent including:
- Australia’s AI Ethics Principles designed to ensure AI is safe, secure and reliable
- Policy for the responsible use of AI in government strengthening the approach to safe and responsible AI through new measures on AI governance
- guidance for agency and staff to understand how to safely and responsibly engage with public generative AI tools
- privacy obligations, outlined in the Guidance on privacy and the use of commercially available AI products and Guidance on privacy and developing and training generative AI models
- the development of an Australian Government AI Impact Assessment Tool to provide a consistent framework for agencies to identify and consider AI risks and impacts in-line with Australia’s AI Ethics Principles
- the establishment of an AI Review Committee to review high risk uses of AI to ensure issues are considered appropriately
- guidance on AI procurement in government and AI contract clauses for responsible AI procurement
- agencies publishing and updating their AI transparency statements outlining their approach to AI adoption
- agency AI Accountable Officials who are accountable for their agency’s implementation of the Policy for the responsible use of AI in government
- the technical standard for government’s use of AI which provides practical guidance for technical specialists and business owners embedding AI in government systems
- the National AI Centre’s Guidance for AI Adoption which sets out 6 essential practices for responsible AI governance and adoption.
Keeping safeguards relevant and effective as the technology evolves
The technology is fast paced. To ensure that the guidance and mechanisms stay up to date there are a number of other pieces in place to support safe and responsible adoption of AI by the APS:
- The AI Delivery and Enablement unit within the Department of Finance will provide a means for considering the implications and assessing how or if the APS AI Plan might need to evolve or change.
- The Digital Transformation Agency will ensure that relevant safeguards continue to be relevant, effective and fit for purpose.
- Foundational training and resources will be provided to all public servants and available through the Australian Public Service Commission (APS Academy) to build capability to use AI responsibly and address gaps.
- By strengthening existing consultation frameworks, there will be meaningful avenues for input so staff are engaged and supported through ongoing change as we together build a more capable and future-ready public service.
- The GovAI ecosystem provides core AI technologies and capabilities that support secure, auditable AI across the APS.
Together, these provide a rigorous and adaptive framework to help ensure that the use of AI by the Australian Government is ethical, moral, legal and people-first. This will allow the government to take a proactive approach throughout AI implementation to give Australians confidence their data is protected. It shares the benefits of this new technology with all Australians.
The Government is also committed to adopting AI in line other commitments, such as the Working for Women Strategy and Closing the Gap targets.
(Please note that the Department of Defence and the National Intelligence Community are exempt from some of these elements but are expected to have their own equivalents in place.)