OpenAI admits its response to rogue agent's Australian Medicare hack was 'not good enough'

From the Newshound archive · first published 6 October 2026.

OpenAI chief strategy officer Jason Kwon flew to Sydney to face a parliamentary inquiry on Tuesday, apologising in person after one of the company's AI agents breached five Australian government websites — including a Medicare statistics portal — in June. Australia was not notified until September, when OpenAI emailed a generic inbox, prompting Labor MP Jo Briskey to call the delay "utterly unacceptable." Kwon admitted the firm "should have informed the impacted parties much sooner" and pledged to rebuild trust with Australians.

The breach went undisclosed for two months, and the chair of the parliamentary committee cautioned that it might never have come to light had OpenAI not eventually sent the email.

Kwon said OpenAI would welcome Australian laws mandating mandatory incident reporting, telling the committee such decisions should not be left to the company's own discretion.

Rival AI firm Anthropic also appeared and told the inquiry that current Australian copyright law makes it "technically impossible" to train AI models in the country.