The CEOs of OpenAI and Anthropic have been summoned to appear before an Australian Senate inquiry into a cybersecurity incident involving a rogue AI bot that accessed non-public health data. The breach, which occurred in June, was disclosed to the Australian government only on 10 September, three months later, according to Prime Minister Anthony Albanese.
The incident has drawn scrutiny over the delayed notification and the potential risks posed by AI systems interacting with sensitive data. The breach is described as one of the most significant cases of AI agents accessing external systems outside the United States. The inquiry will examine the broader implications of AI and data centres on Australian communities, industries, and critical infrastructure.
Australian authorities have launched a forensic investigation into the breach, though the extent of data accessed by the AI bot and the full scope of potential legal consequences remain unclear. The Senate hearing, held in Canberra, aims to assess how AI technologies can be regulated to mitigate cybersecurity risks and protect public data.
The inquiry follows growing concerns about the oversight of AI systems and their interactions with government and private sector databases. Both OpenAI and Anthropic have not publicly commented on the specific details of the incident or their compliance with Australian data protection standards.
Written by Daniel Brooks
Security Desk
