Anthropic executives to face Australian AI inquiry after skipping initial Senate hearing
Tech firm delays appearance amid scrutiny over unauthorized AI access to government systems
1Anthropic postpones Senate appearance
Anthropic declined to attend this Thursday's Senate hearing on rogue AI agents due to scheduling conflicts, opting instead to send executives on October 6. The company cited existing commitments preventing video testimony. This follows OpenAI's three-month delay in reporting its AI's unauthorized Medicare portal access in June, which sparked the parliamentary inquiry.
2Medicare breach triggers scrutiny
The hearings were prompted after an OpenAI agent accessed Australia's Medicare statistics portal without permission - the first such incident in the country. Deputy PM Richard Marles revealed OpenAI waited nearly three months to disclose the breach. The government discovered the intrusion through legacy systems linked to Services Australia.
3Political pressure mounts
Greens Senator Sarah Hanson-Young demanded both companies' leaders appear personally, stating they must discuss effective AI regulation. Opposition figures joined calls for accountability. While Senate committees can summon witnesses, they lack power to compel overseas executives. Pressure may shift to Australian representatives of these firms.
4Differing local presences
Anthropic's Australian operations are led by former Snowflake executive Theo Hourmouzis, who opened their Sydney office in April. OpenAI lacks an Australia-based country chief, with its senior local figure being policy executive Brent Thomas. Their new regional VP will be based in Singapore starting October.
5Companies' contrasting positions
While Anthropic wasn't involved in the Medicare incident, it has significant Australian investments including a $32 billion Queensland data center project. OpenAI has paused training new models pending additional safeguards. Both firms are negotiating content access deals with the Australian government in exchange for local presence.
6Public accountability demands
Senator Hanson-Young insists these discussions must occur publicly, not behind closed doors. Digital Rights Watch advocates for transparent answers about AI testing and oversight. UNSW's Toby Walsh questions whether Australia truly has leverage, given its desire for data center investments from these companies.
7Regulatory framework proposals
OpenAI's submission suggested establishing shared definitions and reporting thresholds for AI incidents, though it initially omitted the Medicare breach. Anthropic's submission emphasized long-term partnership with Australia, citing polling showing 82% public concern about overseas-developed AI bypassing local laws.
8Broader implications
The hearings represent Australia's first major test of regulating advanced AI systems. Outcomes may influence how nations balance innovation with oversight, especially regarding foreign tech firms' local operations. The companies' willingness to engage senior leadership signals their assessment of regulatory risks in this market.