The company told a parliamentary inquiry it would support mandatory disclosure rules.
Anthropic is open to Australian laws requiring artificial intelligence companies to disclose data breaches, company representatives told a parliamentary inquiry on October 6. The maker of Claude said it had found no breaches of Australian government systems after an investigation.
David Orr, Anthropic’s head of safeguards, said an extensive investigation following an incident involving an OpenAI agent had found no intrusions into government systems. He said the company had less visibility into incidents within enterprise customers’ systems. Whether a company reports a breach involving government data is likely determined by internal policy rather than law, he said.
David Masters, Anthropic’s head of policy for Australia and New Zealand, said the company would be open to mandatory disclosure requirements. Australia is preparing AI-specific laws due to take effect next year. In September, the government criticized OpenAI for notifying it of an incident involving the Medicare portal about three months after it occurred.
Copyright was also debated. Technology companies want to use Australian material to train models, while current rules require them to negotiate licensing deals with rights holders. ABC executive Kate Gilchrist said the existing system was adequate and opposed an opt-out model under which content could be used unless a rights holder objected. She said this would shift the burden to creators to monitor use of their material online.
OpenAI representatives were scheduled to appear at the same inquiry later on October 6. The committee is one of at least four Australian inquiries examining AI regulation.