SA's $3M AI inquiry to set guardrails for tech firms by July 2027

For AI developers, the South Australian government's royal commission signals that the self-regulation window is closing. Premier Malinauskas, fresh from meetings with OpenAI, Anthropic, and Apple, has made it clear that only government can ensure technology serves society. With a $3 million budget and a report due by 1 July 2027, the inquiry will likely influence AI deployment requirements, transparency obligations, and the liability framework for machine learning systems, making it essential reading for anyone shipping AI products in Australia.

South Australia's Premier Peter Malinauskas announced on 10 August 2026 the creation of Australia's first royal commission into artificial intelligence, marking a historic step toward binding regulation of the technology. The inquiry, backed by a $3 million budget, will be led by three internationally recruited royal commissioners who will begin work in October 2026 and deliver their final report by 1 July 2027. The announcement follows Malinauskas's visit to the United States, where he met with leading AI firms OpenAI, Anthropic, and Apple—signaling that industry engagement and global perspectives will inform the commission's work.

“

The inquiry, backed by a $3 million budget, will be led by three internationally recruited royal commissioners who will begin work in October 2026 and deliver their final report by 1 July 2027.

The premier framed the commission as a response to what he called a "fork in the road" for Australia, emphasizing both the transformative potential and profound risks of AI. "It is absolutely clear that the use of artificial intelligence is an exciting opportunity to advance the cause of humanity," he said. "But it is also true that, unchecked, artificial intelligence does represent a material risk to the way our society operates, and a risk to the future of work if not thought through carefully." This balanced rhetoric underpins the government's intent to design guardrails that "put people first" rather than allowing technology to dictate societal outcomes.

The terms of reference, to be released in four to six weeks, will define the scope of the inquiry. Malinauskas stressed that they are as critical as the commissioners themselves, and will likely address issues ranging from algorithmic bias and data privacy to employment disruption and infrastructure demands. The premier noted that while data centers were a dominant topic in 70% of his U.S. meetings, he dismissed them as merely an "infrastructure play." The real strategic question, he argued, is how AI technologies are deployed within society—an area where only governments, with their regulatory tools and democratic legitimacy, can establish the necessary frameworks.

The SA initiative stands out in a landscape where most advanced economies are grappling with AI governance. The European Union's AI Act, which came into force in 2025, provides a risk-based regulatory model; the United States has relied on piecemeal sectoral guidelines and executive orders; and Australia's federal government has so far only promoted voluntary ethical frameworks and conducted select committee inquiries. A state-level royal commission—endowed with coercive powers to summon witnesses and compel documents—can produce a comprehensive evidentiary record and detailed policy recommendations. Its report could serve as a blueprint for similar commissions in other states, or push the federal government toward national legislation. Malinauskas acknowledged he had not consulted Prime Minister Anthony Albanese, hinting at potential jurisdictional friction, but also asserting the state's right to act.

The estimated $3 million price tag drew swift comment. Malinauskas called it "chicken feed" compared to the cost to society of failing to address AI's challenges. The commission will likely examine the impact of automation on employment, the adequacy of existing antidiscrimination and privacy laws, and the legal liability of developers and deployers. It may also consider the environmental cost of AI compute, given that the premier had discussions on data center infrastructure. The report's July 2027 deadline gives it roughly ten months to conduct hearings and produce findings—a tight but feasible timeline if well-resourced.