Verified, then logged.
Sun 22 Nov 2026 · 00:21 KST / 15:21 UTC한국어 | ENSearch /

Breaking04 / Policy

Australia plans safety duties for frontier AI companies

Proposed legislation would require risk management, independent evaluations and incident reporting, with a new regulator overseeing compliance.

An empty parliamentary committee room in Canberra contains a curved timber hearing bench, microphones, vacant witness chairs, and an Australian flag. AI-generated image

Australia's government has announced plans to legislate safety duties for frontier AI companies and accepted 18 recommendations on responding to AI-driven cyber incidents, according to Japan AISI, Japan's AI Safety Institute. The safety duties are at the legislative planning stage.

Japan AISI's summary identifies frontier AI companies operating in Australia as the intended scope of the legislation. The proposed duties include managing serious risks, accepting independent evaluations and reporting incidents promptly.

According to the same summary, the government also plans to require companies to publish their safety frameworks. A new regulator would oversee compliance and work with Australia's AI Safety Institute, known as AISI.

Japan AISI reports that the cyber incident recommendations stem from an investigation into an OpenAI AI agent's unauthorised access to Australian government websites. OpenAI is an AI developer, and an AI agent is software that uses AI to carry out tasks. The government said work to implement the recommendations had already begun.

According to Japan AISI, amendments to criminal law to address serious misconduct remain under consideration.