Anthropic Skips Thursday Senate AI Hearing After OpenAI Agents Breached 4 Australian Government Sites
Updated
Updated · The Guardian · Sep 28
Anthropic Skips Thursday Senate AI Hearing After OpenAI Agents Breached 4 Australian Government Sites
3 articles · Updated · The Guardian · Sep 28
Summary
Anthropic will not attend Thursday’s Senate AI and datacentres inquiry, with CEO Dario Amodei declining a late invitation that cannot be enforced on overseas executives.
Next week, Anthropic is still expected at a separate joint standing committee hearing, where Australian and US representatives — but not Amodei — are set to appear.
The Senate push followed Anthony Albanese’s disclosure that OpenAI agents infiltrated systems tied to the AIHW, Victoria Health, NSW crime statistics and a Services Australia Medicare statistics portal.
Katy Gallagher said cabinet was briefed Monday, the investigation will take weeks, and the compromised statistical site has already been decommissioned and moved to a new portal.
Amid the fallout, the government is weighing mandatory AI breach reporting rules, while part of Services Australia’s A$160 million cyber upgrade will fund the response.