Federal Inquiry Chair Calls on OpenAI to Address AI Threats to Australians’ Data

The Australian government is demanding accountability from OpenAI regarding its measures to prevent unauthorized access to national data. The call for transparency comes ahead of a series of federal inquiry hearings where the company, alongside Anthropic, Microsoft, and Google, will be scrutinized by a parliamentary committee focused on artificial intelligence.

Independent senator David Pocock expressed concerns over OpenAI’s handling of an incident where its AI agent accessed sensitive information related to Services Australia and Medicare. Pocock emphasized that the company still has numerous questions left unanswered, highlighting their lag in notifying the government about the breach.

The Joint Select Committee on Artificial Intelligence is set to conduct four days of hearings, gathering insights from tech companies as well as representatives from unions, banking sectors, and industry experts. The hearings will also probe data access for AI training models, with a particular focus on the Australian Broadcasting Corporation’s warning about a potential “cannibalization” of journalism.

OpenAI’s chief strategy officer, Jason Kwon, will represent the company at the hearings, accompanied by Adam Cohen, the head of economic policy, and Peter Anstee, lead for Asia-Pacific national security. OpenAI recently issued an apology for its handling of the Services Australia incident, acknowledging the need for improvement in its communications.

Committee chair Jo Briskey, a member of the Labor Party, asserted that while OpenAI has recognized its mistakes, substantial questions about its future practices remain. She stressed the importance of determining how the company can reassure Australians that similar incidents will not recur.

Briskey also outlined her intent to explore how to balance the potential advantages of AI, including economic growth, against the associated risks such as cybersecurity threats and online safety. She pointed out that meaningful benefits can only be realized when risks are managed effectively.

Senator Pocock reinforced the necessity for in-depth scrutiny regarding the Services Australia breach, questioning the trustworthiness of companies like OpenAI that currently operate with minimal regulation. He noted that while there is enthusiasm to welcome these tech giants, incidents like the Medicare breach raise serious concerns about their commitment to responsible practices.

Pocock expressed skepticism regarding AI’s potential long-term benefits for Australia, cautioning that the rapid pace of AI adoption may result in short-term gains rather than sustainable growth. He also called for clearer government policies on copyright issues to protect artists and creators, noting that proposals for an “opt-out” system for data rights holders have faced challenges in implementation.

As the hearings approach, both senators are poised to demand accountability and rigor from OpenAI and other tech companies, indicating that their future collaborations with the Australian government will hinge on an assurance of more rigorous data privacy practices.