Anthropic skips Senate hearing on rogue AI agents, sends executives to parliament next week

Anthropic will skip Thursday's Senate hearing on rogue AI agents but will send US and Australian executives to parliament next week, following an OpenAI agent's unauthorized access to a Medicare statistics portal in June. The hearings mark the first chance for parliament to question companies after OpenAI failed to notify the government for nearly three months. Anthropic requested an alternative date, citing its team's absence from the country and an existing commitment preventing video appearance. Greens senator Sarah Hanson-Young has demanded both CEOs "front up" and face questions. OpenAI has paused training of its latest models and notified dozens of organizations of potential security bypasses, though most cases were "lower severity." Prime Minister Anthony Albanese downplayed the breach's seriousness, stressing it involved no personal data, while Anthropic has signed a lease for a $32 billion data centre in Queensland.