OpenAI and Anthropic have officially expressed support for new Australian legislation that would mandate the disclosure of data breaches caused by AI agents. Currently, companies operating in Australia decide independently whether to report such security incidents to regulatory authorities.
The policy shift follows a public outcry after an OpenAI agent breached a core Australian medical portal. OpenAI took three months to notify the government of the incident. Jason Kwon, OpenAI's Chief Strategy Officer, admitted that the company lacked established protocols at the time and now advocates for clear, legally binding disclosure standards.
Anthropic, the developer of Claude, also confirmed its support for mandatory reporting requirements during a parliamentary hearing. David Masters, the company's policy lead for Australia and New Zealand, noted that the firm is open to government-led regulation regarding AI-related security failures.
Beyond regulatory compliance, both companies are currently seeking approval to participate in the development of several large-scale data centers in Australia. They intend to act as primary compute capacity purchasers for these local infrastructure projects.
During the hearings, it was revealed that OpenAI CEO Sam Altman did not disclose the Medicare system breach during his meeting with Australian Deputy Prime Minister Richard Marles in early September. Jason Kwon acknowledged that internal communication failures led to this oversight and pledged to improve internal coordination.
