South Australia is taking an unusually expansive approach to artificial intelligence policy: it is putting the technology itself under a royal commission.
Premier Peter Malinauskas announced Monday that the state will launch a $3 million inquiry into the opportunities and risks created by AI, with work set to begin in October and a final report due July 1, 2027. The commission is expected to examine AI’s effects across education, public health, industry, arts and culture, according to reports from The Guardian and News.com.au.
The inquiry arrives as South Australia is already building a formal public-sector AI strategy. The state’s Office for AI has been developing a framework for how government agencies should adopt the technology, including questions of public value, safe deployment and workforce transformation.
That makes the royal commission notable for its scope. Many governments are approaching AI through targeted rules: privacy, copyright, education, procurement, safety or sector-specific regulation. South Australia is asking a larger question about how a general-purpose technology should be absorbed across institutions that were not designed around it.
From rulemaking to systems thinking
The policy debate around AI has often been fragmented because the technology enters organizations in fragments.
A school sees an assessment problem. A hospital sees a clinical decision-support tool. A government department sees automation. A company sees productivity. A creative industry sees questions about authorship and compensation. A worker may see all of those changes as a single employment problem.
Each issue can be regulated separately. The difficulty is that the consequences interact.
If AI changes how work is performed, that affects training. If it changes training, it affects education. If it changes access to expertise, it may affect health care and public services. If it concentrates infrastructure and data in a handful of companies, competition and sovereignty questions follow.
A royal commission creates room to examine those relationships rather than treating AI as a collection of unrelated software products.
That does not guarantee useful policy. Broad inquiries can become abstract, and AI is developing quickly enough that recommendations risk aging before they are implemented. But the structure reflects an important shift in the governance conversation: the relevant unit of analysis may no longer be the individual model or application. It may be the institution being changed around it.
The timing is deliberate
Malinauskas has framed AI as a technological transition with consequences comparable to earlier industrial shifts. South Australia has also been pursuing AI investment and infrastructure while emphasizing responsible adoption, creating an inherent policy tension that many governments now face.
They want the economic upside of AI development while also responding to concerns about labor disruption, energy use, public-sector accountability, education and safety.
That tension is not a contradiction. It is increasingly the central governance problem.
Governments that wait for certainty will likely regulate after organizational behavior has already changed. Governments that move too quickly risk freezing assumptions about a technology whose capabilities and economics are still evolving.
The more useful approach may be to define durable principles around accountability, human authority, transparency and public benefit while leaving room for technical standards to evolve.
Why it matters
South Australia’s inquiry is significant because it treats AI as an operating-system question for society rather than only a technology-policy question.
That is where the debate is heading elsewhere, too. Organizations are discovering that adopting AI is rarely just a software decision. It changes workflows, permissions, staffing, training, decision rights and accountability. Governments face the same problem at a larger scale.
The commission’s eventual value will depend on whether it can translate that broad mandate into practical rules and institutional design. But the premise is increasingly difficult to dispute: AI policy cannot be separated cleanly into technology on one side and society on the other when the technology is being built directly into how institutions operate.