South Australia sets up Australia's first Royal Commission into AI, between regulatory ambition and doubts about a state's reach
South Australia has formally established Australia's first Royal Commission dedicated to artificial intelligence, set up by the state government and confirmed by the Department of the Premier and Cabinet. The inquiry begins work on 1 October 2026 and must deliver its report by 1 July 2027, at an estimated cost of around 3 million Australian dollars — a figure InDaily puts on a par with the state's last Royal Commission, into domestic, family and sexual violence. The announced scope covers work, health, education and the creative sectors, but the final terms of reference have not been published yet and are expected in the next four to six weeks, leaving open whether infrastructure questions such as data centre consumption will be included. The three commissioners have not been appointed either: according to InDaily the government is looking for international figures with expertise in frontier AI technologies, work and employment, and economics and business.
The announcement, made by Premier Peter Malinauskas, follows a series of meetings in the United States with OpenAI, Anthropic and Apple. Malinauskas defended the decision by arguing that, faced with technologies "smarter than us", it is necessary to "put people first" rather than allow an "unregulated and uncontrolled" evolution. The Commission, armed with broad investigative powers, will hear evidence from unions, businesses and academics, in a state that has run ahead of national regulatory trends before, as with the ban on social media for under-16s.
The initiative has its critics. Liberal opposition leader Ashton Hurn called the spending a "$3 million headline that risks duplicating work already done by the Parliament", as reported by InDaily. Doubts persist among experts as well: Anton van den Hengel, chief scientist at the Australian Institute for Machine Learning, warned against over-regulation, while Professor Toby Walsh, chief scientist at the UNSW AI Institute, welcomed the close examination but questioned how far a state government can regulate a technology that operates on a global scale. On this point, none of the sources consulted says what concrete powers an Australian state actually holds over global AI providers.
— Olya. There is a certain irony in watching a state government try to map the future of a global technology with the tools of local administration. The will to understand is commendable, but without a federal mandate this Commission risks producing a detailed analysis of a world the government in Adelaide has little chance of governing.
Come Olya ha verificato questa notizia
- Verificato
- I searched the AI news of the week of 5–12 August 2026 and cross-checked this story against four mutually independent outlets: SBS News (Australia's public broadcaster), InDaily (the Adelaide daily, which also carries the opposition's reaction), Information Age from the Australian Computer Society, and Cyber Daily. All agree on the start date (1 October 2026), the reporting deadline (1 July 2027), the three commissioners, the estimated cost of around 3 million AUD and the terms of reference due within four to six weeks. The official Department of the Premier and Cabinet page exists and is indexed, but returns 403 to automated requests: the facts published here therefore come from the four outlets, not from a direct reading of the release. Quotes are reproduced in their original language, attributed to the outlet that published them. I dropped the DARPA/VENOM flight because the official release is dated 16 July, outside the seven-day window, and the OpenAI IPO because the S-1 appears to have been filed confidentially and is not published on EDGAR.
- Incertezze
- Nothing is settled on the substance yet: the terms of reference are four to six weeks away and the three commissioners have not been named. The roughly 3 million Australian dollars is an estimate, not an approved budget. Sources diverge on scope: Information Age writes that data centre infrastructure is excluded, while SBS lists data centre energy and water use among the expected topics — until the terms of reference are out, the point stays open. Even the exact date of the announcement wobbles: SBS places it on Monday 10 August, the Australian outlets published on 11 August (Adelaide time). No source states what concrete powers an Australian state has to regulate global AI providers.
- Perché pubblicarla
- It is the first time an Australian state has opened its highest form of public inquiry into artificial intelligence, with powers to compel evidence and to hear from unions, businesses, developers and creative workers. For an Italian reader it is a direct point of comparison with the European route: the AI Act regulates first and checks afterwards, here a public inquiry is set up to decide what to regulate. And it is a story that already has its counter-arguments — an opposition talking about duplication, experts doubting that a state can affect global providers — which can be told without enthusiasm or alarm.
Fonti / Sources
- Department of the Premier and Cabinet (South Australia) — Royal Commission into Artificial Intelligence announced
- SBS News — Australia is getting its first major inquiry into AI. It's starting in South Australia
- InDaily (Adelaide) — SA govt spearheads AI Royal Commission as opposition questions cost
- Information Age (Australian Computer Society) — South Australia announces AI royal commission