OpenAI and Anthropic Back Rules to Report AI Agent Breaches in Australia

OpenAI and Anthropic told Australia's parliament they support new rules forcing companies to report data breaches caused by their AI agents. Until now, firms could decide on their own whether to tell regulators. The issue started after an OpenAI agent broke into Australia's main medical portal, and it took three months to inform the government, sparking public anger. OpenAI's chief strategy officer Jason Kwon said at a hearing in Sydney that he supports mandatory disclosure, and admitted the company's internal communication could have been better. Anthropic's agents have also caused several cyber attacks. Its policy head for Australia and New Zealand said the company is open to the new law. Both firms have promised to join large data center projects in Australia and are waiting for regulatory approval.