Submission 8
Department of Finance, Digital Transformation Agency, Australian Public Service Commission and Services Australia
Submitter group: Government.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
Describes the APS AI Plan's mandatory agency safeguards, shared tools, training and service applications. Says adoption includes operational systems, trials and pilots, with benefits requiring assessment in each use case. Services Australia describes staff-support tools and states that AI will not determine statutory entitlements, which remain with authorised human decision-makers.
Key points
Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.
Says individual agencies remain responsible for their own AI adoption and use, within common whole-of-government policies and standards.
Individual Commonwealth departments and agencies remain responsible for their own adoption and use of AI, including the decisions, risks and obligations associated with individual AI use cases, within common policies, frameworks and standards that support consistent adoption across the Australian Public Service (APS).
Describes requirements for in-scope agencies: designate an accountable official, assign use-case accountability, register in-scope uses, assess impacts where required, publish transparency statements and train staff.
Agencies are required to: • designate an AI Accountable Official • assign accountability for individual in-scope AI use cases • maintain an internal register of in-scope AI use cases • undertake AI impact assessments where required • publish and regularly update an AI transparency statement • implement mandatory foundational training on responsible AI use.
States that the AI Review Committee's advice does not displace agency responsibility for the decision, deployment or operation of a use case.
The Committee’s advice does not displace agency responsibility for the decision, deployment or operation of the use case.
Describes GovAI’s onshore hosting and model access as designed to strengthen resilience, preserve choice and reduce reliance on a single provider or technology.
GovAI’s onshore hosting and model-access services are designed to strengthen resilience and preserve choice by giving agencies access to a range of AI models and reducing reliance on any single provider or technology.
Services Australia states AI will not be used to determine statutory entitlements, which remain with authorised human decision-makers.
AI will not be used to determine people’s statutory entitlements. Instead, we are using AI to assist our staff to work more efficiently, improve service delivery, and provide better information and decision support, with all entitlement decisions remaining the responsibility of authorised human decision-makers.
States that more than 100,000 staff have completed foundational responsible-AI training and more than 130 Chief AI Officers have been appointed.
More than 100,000 staff have completed foundational training in responsible use of AI, and more than 130 Chief AI Officers have been appointed across the APS.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Agencies should identify expected AI benefits, establish measures and assess realisation during implementation to inform continuation, adjustment, scaling or reuse.
Agencies should identify the expected benefits of AI initiatives, establish appropriate measures and assess whether those benefits are realised as implementation progresses. This evidence can inform decisions about whether initiatives should be continued, adjusted, scaled or reused, and help distinguish demonstrated public value from anticipated benefits.