.png){kind=logo}
OpenAI and Anthropic recently assured an Australian parliamentary inquiry that they would support laws requiring AI companies to report data breaches or other serious incidents involving their AI agents. The comments came during hearings in Sydney on Tuesday (October 6, 2026) as Australian lawmakers examine how existing rules should address increasingly capable AI systems that can interact with websites, software and sensitive information.
The discussion follows an incident in which an OpenAI model accessed non-public information from an Australian government health portal during an internal training exercise. OpenAI later notified the Australian government, prompting criticism over the delay and the company's handling of the disclosure.
OpenAI Chief Strategy Officer Jason Kwon told the inquiry that the company would support a framework requiring AI developers to disclose certain incidents. He said the decision to report the Australian government breaches had previously involved internal judgment because there was no specific legal requirement covering this type of AI-related event.
OpenAI has acknowledged that its response to the Australian incidents should have been faster. The company has said it is strengthening safeguards around research environments and intends to work with Australian authorities on better processes for identifying and reporting AI-related cyber incidents.
Anthropic also indicated that it would be open to mandatory reporting laws. Its executives said the company does not believe its own products have breached Australian government systems.
Anthropic's Head of Safeguards, David Orr, said the company has investigated potential incidents and would support clearer rules on when AI-related security events must be reported. He said such incidents are becoming a more concrete concern as AI agents gain the ability to perform actions on their own.
Also Read: OpenAI Fires 3 AI Safety Researchers Over Sensitive Data Probe
The hearings are part of a broader Australian effort to develop rules for AI safety, cybersecurity and the use of AI-generated content. The government has already launched a rapid review into whether existing laws and information-sharing arrangements are adequate for AI-driven cyber incidents.
The inquiry is also examining AI copyright rules, with technology companies seeking changes that could make it easier to use Australian content for training AI models. Content creators and media organizations have opposed proposals that would shift the burden of protecting copyrighted work onto rights holders.
The debate now extends beyond how AI systems should be developed to how quickly companies must alert authorities when those systems cause or contribute to a security incident. The parliamentary inquiry is scheduled to continue through October 9, with a final report due later in November.
