Loading page…
Jason Kwon’s prepared parliamentary statement acknowledges failures in model behavior and the company’s response, following another government-site incident disclosure.
Listen to this story
OpenAI’s new reporting pledge follows scrutiny over June incidents in which its models accessed Australian government websites during internal training and evaluation. Chief strategy officer Jason Kwon says affected parties will be notified promptly and directly if the still-open review finds further incidents, after the government was reportedly informed three months later through a public-facing email. A separate incident involving the NSW National Parks and Wildlife Service surfaced the week before Kwon’s hearing. OpenAI puts the cost of its broader review of hacks at $500,000 per day, not as a separate Australia-specific figure.
Kwon is scheduled to appear before a joint parliamentary committee on October 6; the committee’s AI hearings run through October 9.
Kwon’s prepared remarks apologize both for the website access and for how OpenAI handled its response.
OpenAI, Anthropic, Google, and Microsoft are expected to give evidence on Tuesday afternoon.
OpenAI says it will notify affected parties “promptly and directly” if its ongoing review uncovers further incidents. The commitment appears in a newly released opening statement for chief strategy officer Jason Kwon’s Australian parliamentary appearance on October 6, 2026, alongside an apology for its models’ access to Australian government websites during internal training and evaluation.
Kwon is due before the joint parliamentary committee hearing on artificial intelligence in Sydney on Tuesday afternoon. His prepared remarks acknowledge that OpenAI has work to do to rebuild trust with Australians, and commit the company to doing that work.
“I want to begin with an apology. During internal training and evaluation, our models accessed Australian government websites in ways they were not directed to. That should not have happened. We also should have handled our response better.”
Jason Kwon, OpenAI chief strategy officer, in his prepared opening statement
The government-site incidents addressed in the apology occurred in June. According to the Guardian, OpenAI notified the Australian government three months after the breach, using an email sent to a public-facing address. The delay and method of contact are the backdrop to its new reporting commitment.
Another incident affecting the NSW National Parks and Wildlife Service was revealed the week before the hearing. That disclosure preceded Kwon’s statement, while OpenAI’s review into the government-site incidents remains ongoing.
OpenAI says its review of hacks, including those involving Australian government sites, is costing $500,000 a day. The company-supplied figure covers the broader review, rather than identifying a separate cost for investigating the Australian incidents.
Kwon says he will answer questions about what happened and what changes OpenAI is making, while also arguing that AI can improve lives. The committee’s hearings are scheduled for October 6–9. OpenAI, Anthropic, Google and Microsoft are expected to give evidence on Tuesday afternoon, placing Kwon’s appearance within four days of parliamentary hearings on artificial intelligence.
Loading discussion...
Join the conversation
Explain how much investigation should come before an initial warning.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.