The Australian federal government has unveiled a major financial commitment to supercharge its own artificial intelligence capabilities, with a multi-million dollar injection detailed in the latest budget update.
Budget Breakdown: Funding the AI Ambition
The Mid-Year Economic and Fiscal Outlook (MYEFO) revealed a $225 million investment over four years from 2025-26 to advance the government's AI plans. An additional $400,000 per year is allocated from 2029-30 to sustain the initiative.
The lion's share of the funding, $166 million, will be jointly managed by the Department of Finance and the Digital Transformation Agency (DTA). This money is earmarked to expand the existing GovAI platform and to design, build, and pilot a new secure AI assistant named GovAI Chat.
Finance Minister and Minister for the Public Service, Senator Katy Gallagher, launched the government's AI adoption strategy for the Australian Public Service (APS) in November 2025. Senator Gallagher stated the goal is for every public servant to have access to secure, generative AI tools directly from their work computers.
Training, Safety, and Expert Oversight
Beyond the core platform, significant funds are directed towards supporting functions. A sum of $22.1 million will go to the Australian Public Service Commission and the DTA to develop AI training programs and new staffing approaches to help agencies adapt to changing roles.
To address risks, $7.7 million over four years is provided to the DTA to establish an AI review committee. This expert panel will advise on high-risk government AI uses and ensure responsible deployment across the APS. Furthermore, the Department of Industry, Science and Resources will receive $30 million over four years to create a new AI safety institute, focusing on protecting Australians from AI-related risks.
Industry Reaction: A Positive First Step with Caveats
Michael Gately, CEO and co-founder of Canberra-based AI firm Trellis Data, described the funding as a positive domestic initiative but noted it lags behind international investment scales. He cautioned that funding training alone does not guarantee productivity gains.
"There might still be quite a wide proficiency spectrum ... everyone gets trained, but not everyone will really know how to benefit from that," Mr Gately said. He stressed the need for clear key performance indicators (KPIs) to measure genuine benefits, such as reducing unemployment or speeding up application processing, rather than superficial time savings.
On the complex issue of AI safety and governance, Mr Gately suggested that managing risks might require public-private partnerships. Sovereign companies could provide the deep technical knowledge needed to understand AI models at a "molecular level," while the APS focuses on building public trust.
The funding announcement coincides with a shift in the government's National AI Plan, moving away from implementing hard rules for risky AI in the short term. Instead, the plan will rely on existing laws and expert advice, following feedback from the business sector and the Productivity Commission.
