Sydney (AFP) – Australia announced on Tuesday that it would impose strict prudential measures on companies specializing in artificial intelligence (AI) wishing to operate in the country, after the hacking of government platforms by an OpenAI model.

Published on: 10/11/2026 - 12:20

A new AI regulatory authority, responsible for implementing measures aimed at regulating this sector, will be created, indicated Prime Minister Anthony Albanese.

His government has positioned itself at the forefront of a fight to regulate tech, for example by banning many social networks for under-16s since the end of 2025 - even if the measure is largely circumvented, according to a study published in August.

“We cannot turn a blind eye and act as if nothing had happened. But what we can do is ensure that this technology (AI) is put to the service of Australia's interests,” the head of government told the press.

New safeguards will include mandatory reporting of AI risks and independent assessment of cutting-edge models.

“Australia is a pioneer in this area and we want to ensure we take full advantage of the potential benefits of AI,” said the Prime Minister. “But we have to make sure we get it right, and societal buy-in is absolutely essential,” he added.

Concerns are growing globally about the capabilities of advanced AI tools, after a series of self-initiated hacks by models from OpenAI and its competitors Anthropic and Google.

Hacks

In June, an autonomous OpenAI agent infiltrated an Australian government website and accessed files from the Health Statistics Service. Canberra accused the American company of only alerting it in September, by email to a generic email address intended for the general public.

In total, four Australian government platforms were targeted by the ChatGPT creator's AI.

For Defense Minister Richard Marles, "it is important that AI models are subject to an independent evaluation" and that "risk management" in this area be strengthened, he told the press.

Australia, of which large areas are uninhabited, is a place of choice for the construction of “data centers”, these installations necessary for the functioning of AI. OpenAI is particularly involved in the construction of a huge data center on the outskirts of Sydney.

But voices have been raised about the environmental impact of these sites which are particularly energy and water intensive.

And Australia's arts sector has accused Anthropic and other AI companies of seeking to relax copyright laws so that chatbots can be trained on local songs and books.

© 2026 AFP

The content you are trying to access does not exist or is no longer available.