IT Home News on the 6th of the 10th month: According to Reuters, OpenAI and Anthropic told the Australian Parliament on Tuesday thatsupport introducing relevant regulations that would require companies to report data breaches caused by their AI agents. Both organizations also admitted that notifying regulators is currently entirely at companies' own discretion.
The statements came in response to an incident in which an agent under ChatGPT developer OpenAI breached Australia's core medical portal, and the organization took three months to notify the Australian government, triggering strong public outcry.
Currently, both OpenAI and Claude's developer Anthropic have committed, as core compute purchasers, to participate in several large data center projects planned by local Australian developers, and the two organizations are awaiting relevant regulatory approvals.
OpenAI's Chief Strategy Officer Jason Kwon stated at a hearing held in Sydney: "We support establishing a mandatory disclosures mechanism."
Regarding OpenAI's model breaching the Australian medical portal and three other government websites, Kwon explained to the parliamentary inquiry committee: "At the time we were working out our response processes and trying to develop a set of applicable standards."
He added: "This is precisely the role that legal regulations can play. Elected representatives such as legislative bodies need to take on more decision-making responsibilities, so that standards don't have to be set unilaterally by us companies alone."
AI agents under Anthropic had previously also been involved in several cyberattack incidents. David Masters, the company's head of policy for Australia and New Zealand, likewise said at the hearing that the company is open to Australian legislation requiring AI companies to proactively disclose data breaches.
IT Home noted that Australian Deputy Prime Minister Richard Marles previously said that when he met with OpenAI co-founder and CEO Sam Altman at the beginning of month 9 (early September), the latter did not mention the breach of Australia's national health insurance system (Medicare).
In response, Kwon defended to the parliamentary inquiry committee thatAltman was unaware of the breach at the time, even though other departments within the company had already been aware of the situation. Kwon admitted: "I acknowledge that the internal information-sharing process on this incident within the company could have been better. We want to make sure such problems never happen again."