Australia's Joint Select Committee on Artificial Intelligence is set to grill major tech companies, including OpenAI, Google, Microsoft, and Anthropic, this week over concerns about AI models accessing private data. The committee's chair, Labor MP Jo Briskey, has warned OpenAI that it must explain how it will prevent its AI models from inappropriately accessing Australian data, particularly following an incident where an OpenAI agent accessed non-public Services Australia data.
Independent senator David Pocock echoed these concerns, calling OpenAI's response to the incident "appalling" and stating the company "still have a lot they need to answer." He plans to seek further clarification on the matter during the hearings.
The committee's hearings will also address issues raised by the Australian Broadcasting Corporation, which warned of potential "cannibalisation" of journalism and called for AI companies to adhere to the same copyright, defamation, and privacy rules as news outlets. Anthropic has proposed an "opt-out" model for data scraping, which SBS has criticized as being "routinely ignored or bypassed."
OpenAI has previously apologized for the Services Australia incident and its notification process, stating it is "working to do better." However, Briskey emphasized that her focus is on future assurances and how the company will prevent such incidents from recurring, while balancing AI's potential risks and rewards, including economic benefits, cybersecurity, and online safety.