Defence Australia puts data at core of national security

Defence Australia is positioning data and artificial intelligence as core elements of national security, as sovereign AI capabilities become equally important to military readiness as traditional defence assets.

Its One Defence Data program is helping transform fragmented information environments into a connected data ecosystem designed to support faster and more informed decision-making.

The importance of data has been reinforced through successive government strategic reviews and investment programs, with information increasingly viewed as a source of operational advantage.

Speaking on a panel at the Gartner Data & Analytics Summit in Sydney, Defence Australia CTO, Nasa Walton, reinforced this sentiment.

"For centuries, military might has been measured in the volume of ships and planes," Walton said.

"What has been missing is how data has become a military strength and a military might. Especially over the last few years, with the onsets of sensors and information that we have in our capability, data has become a strategic advantage for defence."

Defence's challenge has not been centralising all information into a single repository but creating a federated architecture that allows data to be shared and trusted while remaining within operational systems, where appropriate.

Its assets, such as aircraft, land vehicles and vessels, generate thousands of individual data points during operation, making wholesale replication impractical. Instead, the One Defence Data environment is designed to connect information sources and deliver relevant data to decision-makers when required.

The program is intended to support command and control functions by ensuring information is secure, trusted, sovereign and available at speed, and iis increasingly being used to combine structured and unstructured data while enabling information to be reused across multiple applications and analytics systems.

Organisations across sectors are facing similar challenges around connecting data, removing silos and ensuring technology investments delivered measurable outcomes

The Defence approach reflects broader trends, across industries such as financial services, emergency service and logistics, where organisations increasingly rely on connected data environments to make decisions in real time, according to Cloudera ANZ CTO, Vini Cardoso.

"The role of the CTO these days is not just about selecting technology, not just about provisioning capacity, but making sure that every single investment that is made is measurable towards a proper outcome," Cardoso said.

As AI adoption accelerates, both executives argued that governance frameworks are becoming critical to successful deployment.

Defence has focused on using its own data assets to train and refine AI systems rather than relying solely on externally trained models. Walter argued that workforce capability remains just as important as technology investment, with successful AI initiatives dependent on staff who understand data, algorithms and operational requirements.

The sovereignty of its data is of paramount importance. 

While public AI services have helped drive experimentation, Defence requires significantly higher levels of assurance around data security, model behaviour and information integrity.

Defence must not only protect sensitive information but also understand how misinformation could be introduced into AI systems and influence decision-making.

Intellectual property and copyright issues emerging from AI-assisted software development is another factor to consider, with organisations needing to remain aware of how commercial models are trained and what material may have been incorporated into them.

The risk of being left behind by AI

Organisations risk playing catch up if they fail to embrace artificial intelligence, but moving too quickly without proper governance could prove equally damaging.

Companies operating in regulated industries are increasingly seeking ways to deploy AI while retaining control over their data and intellectual property.

"When you bring the models under your control, that can be in your data centre or in your cloud of choice, but it's ring-fenced to your needs in your private network," Cardoso said.

The approach is designed to address concerns around data sovereignty, compliance and intellectual property leakage, particularly as organisations begin integrating AI into core business processes.

Recent well-publicised incidents where employees inadvertently exposed sensitive, proprietary information through public AI tools are a warning for companies to maintain appropriate governance.

"No one wants to have a situation like that," Cardoso said.

While concerns about data exposure continue to drive interest in private AI deployments, he warned that excessive caution could also create risks.

"Organisations that don't embrace AI will inevitably be left behind, and you're going to become insignificant," Cardoso said.

At the same time, businesses rushing to deploy AI without governance frameworks, clear strategies and operational controls could encounter significant problems.

Many organisations continue to struggle with fragmented and poorly governed data environments, limiting their ability to move AI initiatives beyond pilot projects.

AI should not be viewed as a shortcut for solving longstanding data management challenges. However, AI itself can play a role in improving data governance and visibility.

Cloudera's own data lineage capabilities, for example, use AI to help organisations understand where data originates, how it is transformed and where it flows across the business.

A lack of visibility into enterprise data remains a significant challenge, with many organisations unable to fully map data stored across departmental applications, shadow IT systems, shared drives and employee devices.

Beyond governance concerns, cost management is emerging as another obstacle as enterprises expand AI usage.

Some companies have significantly underestimated the financial impact of large-scale AI deployments, particularly around inference costs and token consumption.

"I know organisations that have a yearly budget and they spend their budget in the first two months," Cardoso said.

As AI adoption broadens across workforces, controlling usage and monitoring expenditure is becoming a growing concern for technology and finance teams alike.

As well as the technical challenges, cultural change could cause friction, too.

While modern platforms can connect hundreds of systems and make data more accessible, organisations often struggle to overcome internal resistance to sharing information.

Successfully scaling AI requires not only technology investments but also organisational change, with businesses needing to establish governance frameworks that encourage responsible data sharing while maintaining appropriate safeguards.