All Australian Public Servants to Get AI Chatbot Access by 2026
Govt AI Chatbot for All Public Servants by 2026

In a major technological advancement for the Australian Public Service, every federal public servant will gain access to an in-house generative artificial intelligence chatbot within the next year. The ambitious rollout forms the cornerstone of the government's newly released AI strategy, announced by Public Service Minister Katy Gallagher.

Comprehensive AI Rollout Across Government

Senator Gallagher unveiled the Australian Public Service's AI plan at the Government Innovation Showcase in Canberra, emphasising its potential to transform public sector operations. The minister predicted that no single area of government would remain untouched by AI in the coming years, highlighting both the opportunity and responsibility this technological shift represents.

"If that's the case, and I'm pretty certain it will be, the responsibility falls to all of us to guide and shape how that happens," Senator Gallagher told attendees comprising public servants and industry representatives.

The centrepiece of this initiative is the expansion of GovAI, the government's secure AI platform that has been undergoing testing since July 2025. Under the new plan, GovAI will be rolled out to the entire public service with full operational status expected by April and complete implementation by the end of 2026.

Secure AI Tools and Governance Framework

The enhanced GovAI platform will feature GovAI Chat, providing public servants with secure, in-house AI tools accessible directly from their work computers. Every public servant will receive comprehensive training and guidance on using these tools safely and responsibly, addressing potential concerns about AI implementation in government services.

To ensure proper oversight and accountability, the government will establish several key governance structures. Every government agency must appoint an SES-level chief AI officer by July 2026, responsible for driving AI adoption and ensuring compliance within their organisation.

Additionally, the Department of Finance will create an AI delivery and enablement (AIDE) team to coordinate AI adoption across the public service and accelerate priority use cases. Senator Gallagher emphasised that agency secretaries and their senior leadership teams would bear ultimate responsibility for championing technological change within their portfolios.

Learning from Past Technology Failures

Acknowledging the shadow cast by previous government technology failures, Senator Gallagher directly addressed the robodebt scandal, noting that it frequently comes to mind when people consider technology adoption in the public service. While emphasising that robodebt wasn't about AI, she stressed the importance of robust governance and oversight.

The AI plan centres on three core pillars: trust, people, and tools. To maintain public confidence, the government will establish an AI review committee to provide whole-of-government oversight and assess all high-risk AI applications across the public service.

Senator Gallagher clarified that the initiative focuses on providing general access to AI tools rather than automating critical decision-making processes. "It's not about bringing AI in to decide that someone gets a payment or not," she stated. "That is not where we're at, and being clear about where we expect humans to be making decisions is very clearly set out in the policy frameworks."

The minister described the expected costs of implementation as "modest" considering the potential productivity improvements and enhanced work processes. These expenses will be detailed in the upcoming mid-year budget update.

Senator Gallagher highlighted several potential applications for AI in government services, including automatically reuniting Australians with lost superannuation or unclaimed Medicare benefits, anticipating significant life events to connect people with eligible services, and detecting fraud and organised crime by identifying patterns across massive government datasets that would be invisible to human teams.