SA hits AI with a $3m royal commission — an Australian first

Australia's first major public inquiry into AI isn't federal — it's coming out of South Australia. Premier Peter Malinauskas has called a royal commission into AI, the country's heaviest-duty investigative instrument: compelled testimony, summoned documents, the works. It kicks off in October 2026, costs roughly $3 million, and must report back by 1 July 2027.

The scope is deliberately wide — the future of work and job disruption, AI's energy and water consumption, health and education, the creative industries, and what regulation should actually look like. Three commissioners get named in the coming weeks; terms of reference land in four to six weeks. Evidence will be pulled from unions, business, academics and the model developers themselves. The timing isn't accidental: Malinauskas announced it after a US trip that included meetings with OpenAI, Anthropic and Apple.

The obvious objection is jurisdictional. A single state can't regulate frontier labs, and critics warn that a patchwork of state rules could hobble Australian AI firms without touching anyone abroad. But UNSW's Toby Walsh and others read it as a forcing function — a formal, evidence-taking process that drags a national debate into the open while Canberra is still deciding whether it has one.

Why it matters: when a government reaches for its most powerful legal tool, the question shifts from "should we regulate AI?" to "on what evidence?"