Redefining sovereignty in the age of AI: control and opportunity
Can we, or should we, aim to replicate traditional models of sovereignty in the virtual world?
Can we, or should we, aim to replicate traditional models of sovereignty in the virtual world?
Sovereignty has long been associated with a state鈥檚 role in providing collective security - protecting its citizens from both internal instability and external threats. As our lives extend increasingly into the digital realm, we are being forced to reconsider what sovereignty means and how it functions in this new domain.
Digital sovereignty, particularly in relation to artificial intelligence, challenges the conventional boundaries of state control. Can we, or should we, aim to replicate traditional models of sovereignty in the virtual world?
This question was at the heart of a recent discussion dinner at The Boathouse in Canberra, hosted by the 麻豆社madou Institute for Cyber Security and the 麻豆社madou AI Institute.
Held under the , a diverse group of cyber security leaders, technologists, strategists, and policy makers gathered to explore what sovereignty in AI might look like for Australia - and whether it remains the right concept at all.
One thing quickly became clear: sovereignty in AI means different things to different people. For some, it starts with data - if we don鈥檛 own, label, and store it in a way that reflects Australian values and priorities, we lose the foundation on which AI systems are built.
Others questioned whether sovereignty is really about control at all: is access different from ownership? Who holds the kill switch?
Underlying these questions is a deeper tension: if we lean too far toward control, we risk isolationism. But if we fail to invest in sovereign capability, we become passive recipients of other nations鈥 priorities and values.
Somewhere between these poles is the possibility of balance - a model of digital sovereignty that enables us to shape our future without closing ourselves off to global collaboration.
To achieve this balance, we need to be clear-eyed about the risks, realistic about our resources, and bold in our ambition. Not every element of the AI ecosystem must be sovereign.
As one guest noted, 鈥淎I is scary to many, especially as we encounter an ageing population here in Australia. We have to deal with the idea that AI is this magical beast that can do everything. It can鈥檛.鈥
Instead, we need to ask: where can Australia lead? Where do we have an edge - whether in infrastructure, regulation, or research - and how can we use partnerships to supplement our strengths rather than undermine them?
Several participants reflected that other countries seem more willing to experiment and accept risk. Australia鈥檚 comparative caution could hinder its ability to compete unless we begin to act with greater urgency.
Ultimately, sovereignty in the digital world must serve both protection and freedom.
It must ensure that Australians are secure - not only from technological harms, but also in their ability to fully participate in digital life. That means sovereignty cannot be an exercise in control alone.
It must also be an enabler of trust, innovation, and resilience.