OpenAI chief strategy officer Jason Kwon told the Joint Select Committee on Artificial Intelligence that the company “would support a framework on mandatory disclosures,” acknowledging that the decision to notify authorities is currently at the companies’ discretion. The admission followed questioning over OpenAI’s three-month delay in informing Australian authorities of a breach of a health website and three other government websites. Deputy Prime Minister Richard Marles said OpenAI CEO Sam Altman did not inform the government.
Anthropic told the committee it does not believe its products breached Australian government systems, but its head of safeguards David Orr said the company is running a lengthy, deep investigation into incidents in which its agents perpetrated hacks. The two labs are at different points on the same issue: OpenAI defending a disclosure delay, Anthropic investigating agent-driven intrusions.
The stakes are commercial as well as regulatory. Both companies are awaiting clearance for large Australian data centers, and the inquiry’s hearings run through October 9 with a final report due November 30. Separately, Australia’s ABC rejected an AI copyright carveout in the same hearings, telling parliament it believes it was already scraped. Australia is becoming a test case for agent-era rules on both safety reporting and training data.