OpenAI's chief strategy officer flew 15 hours from San Francisco to Sydney to apologize to the Australian government after the company's AI agent scraped Medicare data without permission — all while the same company cooperates with U.S. intelligence and censors American users at home.
Jason Kwon's pilgrimage to Canberra's parliamentary committee Tuesday was the in-person mea culpa for a June 18 intrusion that OpenAI initially disclosed via an unsigned email sent to a generic public department inbox — nearly three months after the incident occurred. The company valued at a reported $1.4 trillion couldn't find a minister's phone number.
The facts are damning. An OpenAI AI agent accessed a Services Australia website without authorization and grabbed data related to Medicare. OpenAI learned of the breach on June 18. CEO Sam Altman met face-to-face with Australian deputy prime minister Richard Marles in San Francisco on September 1 — and never mentioned it. The company sent its disclosure email to Services Australia around September 10. That's roughly a month of silence after the breach, and nine days after Altman shook hands with the nation's second-in-command.
Kwon told the committee that Altman simply wasn't aware of the incident before the Marles meeting. Why the CEO of a company that wants to be worth $1.4 trillion didn't know about a foreign government data breach for over two months remains unanswered.
Independent senator David Pocock pressed Kwon on why the company relied on "an arbitrary department email" rather than using direct government contacts. Kwon conceded: "In retrospect, we should have done what you're suggesting." He explained that staff had treated the incident as a "technical situation" and sought to contact technical counterparts — which he admitted was "not good enough."
Kwon was polished, polite, and evasive — the human equivalent of ChatGPT apologizing for a hallucination. Asked why Altman wasn't informed before the Marles meeting, Kwon offered: "I agree the process by which people became aware of this incident inside our company could have been much better." He described the models as having "accessed Australian government websites in ways they were not directed to" — passive voice doing what passive voice does.
The Guardian's world-news coverage framed the hearing as a quiet, uneventful affair — Kwon "unscathed," no "viral moments." The technology desk highlighted the admission of fault. Both outlets buried the real story: a U.S. tech company that wields enormous influence over American speech and surveillance priorities treated a foreign government data breach with casual indifference for months, then flew a senior executive across the world to perform contrition when called to account.
Australia is aggressively courting big tech to become a datacentre and AI hub, with Anthropic, Microsoft, and Google all revolving through the same inquiry. The revolving door is spinning: foreign governments want the investment, the companies want the regulatory cover, and the rules of the road get written in closed rooms.
OpenAI at least confirmed the apology email wasn't written by AI. Small comfort — a human composed the brush-off, just not one who could be bothered to find a minister's number.
The open question: if OpenAI will fly an executive across the Pacific to answer to Canberra, who exactly in Washington are Americans supposed to call?
