14 /
Perspective
The rise of generative technology presents both an opportunity and a reckoning for Australia’s public sector. As agencies explore its potential to transform digital services and internal operations, the conversation has changed, with less focus on technical capability and more on, sovereignty, accountability, and public trust.
At a recent Melbourne roundtable hosted by DataStax and Trideca, senior leaders from government and industry agreed that the goal is not to avoid AI but to adopt it in a way that supports innovation without compromising integrity. This article highlights key outcomes, with a focus on the needs of the public sector.
From Trials to Trusted Systems
Participants confirmed that Australia’s public sector has shown strong interest in this space. The federal government recently led the world’s largest, whole-of-government deployment of Microsoft Copilot. Yet many pilots have since been slowed or paused due to risk, accuracy, and public scrutiny concerns.
Government departments operate under unique pressures. A mistake by a chatbot, a misinterpreted policy, or a data breach can carry significant legal and reputational costs. Public trust is not optional; it’s foundational.
Sovereign Infrastructure Builds Confidence
A key insight from the roundtable is that sovereignty isn’t just about where data is stored; it’s about operational control and explainability.
With platforms like DataStax Astra DB, agencies can implement retrieval-augmented generation (RAG)that only uses approved data, with full audit trails for every output. Crucially, if a reliable match can’t be found, the system can respond with “no answer”—a vital safeguard in areas like regulation, policy advice, and citizen services.
This transparent, modular approach is especially effective for document classification, structured correspondence, and high-volume inquiries where precision is essential.
Enabling Innovation with Guardrails
One participant noted: “Good guardrails are like the lines on a sports field—they don’t stop play, they keep it fair.”
That’s the goal of AI governance. Rather than restrict innovation, policies should embed privacy, security, and fairness directly into the process. Some government agencies now use natural language filters to scan outputs for tone and privacy issues before they’re delivered.
Trideca works with public sector teams to design responsible, contextual solutions that improve efficiency while aligning with human needs and community expectations.
Trust Must Be Earned
Conversation also focused on how Australia ranks near the bottom globally for public trust in AI–42nd out of 47 countries, according to a studyby KPMG and the University of Melbourne.
This means AI systems must be transparent, human-centred, and designed to assist public servants—not replace. From caseworkers to policy teams, the role of AI should be to augment judgment, not obscure it. People need to understand not just what a system says, but whyit says it.
Controlling the Data That Matters
Generative systems rely heavily on data. In the public sector, control over that data is non-negotiable. Without it, agencies face risks around governance, legal exposure, and reliability.
To build public trust in generative AI, systems must be explainable, adaptive in model selection, and able to cross-check outputs for accuracy and authority.
Participants discussed how Sovereign AI ensures models are run within Australia under national laws like the Privacy Act and Consumer Data Right. This protects against external threats, such as offshore outages or foreign subpoenas, and builds local capability.
When models are trained on Australian data and tailored for public needs, results become more accurate, explainable, and relevant.
One participant highlighted that sovereign systems must also allow for human decision points, ensuring oversight where it matters most.
Laying the Groundwork for Future Services
Done right, generative tools can reduce wait times, improve access, and help departments make better, faster decisions. But to get there, agencies and their leaders must modernise legacy systems, enable real-time data use, and build skills across teams.
The focus should be on building trustworthy, effective systems that serve the public, beyond the hype of emerging technology.
With trusted partners like Trideca and platforms built for transparency and control, Australia’s public sector can lead the way, creating digital services that are secure, accountable, and truly fit for the future.
We’ve listed several key points from our recent roundtable above. Let us know if you’d like to attend any of these robust conversations in the future.
The event was a result of the partnership between Trideca and DataStax (an IBM company)
