Top OpenAI boss on hack apology tour 'can't explain' Australian AI distrust

Chief strategy officer Jason Kwon of OpenAI visited Australia following a hack of Medicare statistics, yet he struggled to explain why Australians are among the most distrustful of AI globally. During a parliamentary inquiry, Kwon pledged to rebuild trust but had no answer for the high distrust levels. OpenAI previously praised Australia's adoption of ChatGPT, but Kwon offered no insight into why a majority of Australians do not trust AI, more than in most other countries.

While OpenAI now alerts staff to model misuse during training, Kwon acknowledged the company should have disclosed the Medicare hack to the Australian government sooner, as they identified and reported the NSW Parks and Wildlife breach last week. Kwon also revealed that CEO Sam Altman was unaware of the breach despite staff discovering it weeks earlier.

In Australia, Kwon will interview individuals for OpenAI's promised taskforce to guide its response to the hack. Meanwhile, rival AI company Anthropic supported a proposal by the federal government's Office of AI requiring AI developers to report serious safety incidents, stating it is finalizing a deal to allow Australia's AI Safety Institute to independently test its models.

Australia's current copyright regime was cited as an obstacle, but Anthropic stated it had not signed any licensing deals in Australia. Musician, writer, and voice actor groups rejected calls to loosen copyright laws to attract AI investment, accusing companies of using their work without permission. ARIA chief executive Annabelle Herd warned that AI deals would threaten Australia's music industry, as artists would be "roadkill" in the rush to AI deals.

Both Google and Microsoft appeared before the committee, urging a less catastrophic approach to regulating AI, suggesting that Australia should build on its existing laws and adopt industry standards that work across borders.