Australia is preparing stricter artificial intelligence regulations after an OpenAI agent accessed a government health portal without authorization. The incident has raised concerns about AI security, particularly when autonomous systems interact with sensitive government infrastructure.
Prime Minister Anthony Albanese announced the proposed measures on Tuesday. Under the new framework, AI companies will have to report cybersecurity incidents and unexpected behaviour. Furthermore, the government will consider criminal law changes to address serious misconduct involving AI systems.
Albanese also confirmed that his government had accepted all recommendations from a rapid review. Authorities launched the review following the incident involving OpenAI’s AI agent.
OpenAI Agent Accesses Australia’s Government Health Portal
The incident involved an OpenAI agent accessing Medicare’s medical statistics portal without authorization. The AI system was researching public healthcare spending when the incident occurred.
Medicare serves as Australia’s universal health insurance programme. The government disclosed the security breach in September and subsequently ordered a rapid review.
OpenAI said it became aware of the incident in August. The company described the access as unintentional and provided details of its internal investigation.
According to OpenAI, investigators found no evidence that the AI agent accessed patient records. However, the incident prompted questions about the safeguards surrounding autonomous AI systems.
It also highlighted potential risks when AI agents interact with government platforms and sensitive digital infrastructure.
Australia Considers Mandatory AI Incident Reporting
Australia plans to introduce mandatory reporting requirements for artificial intelligence companies. These rules will cover cybersecurity breaches and unexpected behaviour by AI systems.
Companies will have to report incidents involving AI systems that act against their intended instructions. Consequently, developers and operators will face greater responsibility for monitoring their technology.
The government will also examine possible changes to existing criminal laws. These changes would address serious misconduct involving artificial intelligence systems.
Albanese said existing legislation must keep pace with rapidly developing technology. Therefore, the proposed measures aim to strengthen accountability as AI capabilities continue to expand.
New AI Safety Regulator to Monitor Technology Companies
Australia’s centre-left Labor government will establish a dedicated AI safety regulator. The proposed authority will oversee companies developing and deploying artificial intelligence systems.
Its responsibilities will include monitoring security controls throughout AI development and deployment. Additionally, the regulator will identify serious risks and conduct independent evaluations of AI models.
The authority will also have powers to impose significant penalties on companies violating AI safety requirements.
These mandatory protections will form part of the government’s proposed artificial intelligence legislation. Canberra plans to introduce the legislation in Parliament in 2027.
The move signals a broader effort to strengthen oversight of AI companies. It also reflects growing concerns about the risks associated with increasingly autonomous systems.
OpenAI and Anthropic Support Mandatory AI Reporting
The proposed regulations come amid disagreements between Australia and major US technology companies. Canberra previously rejected requests from OpenAI and Anthropic for exemptions from copyright laws covering AI model training.
Instead, the government requires both companies to negotiate licensing agreements with Australian copyright holders.
Despite these disagreements, the companies have expressed support for mandatory security reporting. Last month, OpenAI and Anthropic told Australia’s Parliament they would support reporting requirements for security incidents involving their AI agents.
Both companies acknowledged that reporting such incidents currently remains at their discretion. However, Australia’s proposed legislation would make these disclosures compulsory.
The planned framework would introduce stronger government oversight of AI development and deployment. It would also establish clearer reporting responsibilities when autonomous systems experience security failures or unexpected behaviour.
Australia’s proposed legislation is expected to reach Parliament in 2027, potentially changing how AI companies manage security risks and report incidents.
For the latest updates, visit and follow The Truth International website (www.thetruthinternational.com) and subscribe to the YouTube Channel.
