Australia wants stricter rules for AI companies after hacker attack

After hacker attack Australia announces stricter rules for AI companies As of: November 10, 2026 • 1:00 p.m

An AI security incident recently caused a shock in Australia - now the government is reacting: the rules for AI developers are to be tightened. Changes are also being examined in criminal law.

Australia wants to tighten security regulations for AI developers. The government is responding to a serious hacker attack by an AI agent on a government website in June.

Following an investigation into the incident, 18 recommendations were made, which Prime Minister Anthony Albanese said the government plans to fully implement. “Our task is to ensure that we shape the development of artificial intelligence - and not that artificial intelligence shapes us,” said the head of government.

There were repeated security-related incidents. This is how AI systems tried to get into other systems. more

Specifically, an AI supervisory authority is to be set up that can monitor security controls and impose drastic sanctions in the event of violations. In addition, possible changes to criminal law should be examined in order to better punish serious misconduct in connection with artificial intelligence.

The new requirements are intended to ensure that AI companies identify possible dangers early, secure their systems better and act transparently when problems arise. It should become mandatory to report security incidents in the area of ​​AI.

Control by independent experts and the publication of security concepts should also become mandatory. “The law needs to stay ahead of, and in some cases catch up with, a very rapidly changing world,” Albanese said.

The focus of the conference is the question of how Germany and Europe can remain competitive in AI. more

The measures follow a security incident in June: an autonomous AI agent from the company OpenAI hacked an Australian government portal without authorization, which, among other things, provides statistics about the country's national health insurance.

The web application had denied the agent access several times - but he ignored this and finally found a way into the system himself. He was also able to access non-public files. However, according to current findings, no personal information is said to have been accessed.

While researching statistics, an AI agent hacked into a health insurance portal. more

OpenAI only informed Australian authorities about the incident in September - via email to a public mailbox. Albanese described the company's behavior as unacceptable. The incident has heightened tensions between Australia and US technology companies.

The government had previously denied ChatGPT maker OpenAI and competitor Anthropic copyright exemptions for training their models. However, both companies told the Australian Parliament last month that they would welcome legislation requiring reporting of data breaches caused by their AI systems.

AI companies are interested in Australia, for example to build data centers. An attack on a government portal raises questions. more

Australia sees itself as a pioneer in AI regulation

Australia's government sees itself as a pioneer in the regulation of highly developed AI systems and speaks of a "world-groundbreaking" project.

The comprehensive AI law with mandatory security and protective measures is expected to be introduced into parliament at the beginning of 2027.

This topic in the program: tagesschau.de | Australia wants tough penalties for misuse of AI | 11/10/2026 | 11:04 am