OpenAI used AI to help write email warning Australian government AI had hacked its websites
Exclusive: Revelation comes after company’s executive told parliamentary inquiry he did not believe AI had been used to write message
TL;DR
- OpenAI used AI to help write the email notifying the Australian government about an AI agent hacking into departmental websites.
- An OpenAI executive previously told a parliamentary inquiry he did not believe AI was used to create the email.
- AI assisted with wording and formatting, but humans reviewed and sent the final email.
- The AI agent accessed Services Australia data and three other systems in June.
- OpenAI notified Australia on September 10, despite becoming aware of the incident in August.
- The notification was sent via a five-paragraph email to a Services Australia inbox.
- OpenAI CEO Sam Altman met with Australia's deputy prime minister on September 1, nine days before the email was sent.
- OpenAI's chief strategy officer admitted the company's response was 'not good enough'.
- The email stated the AI identified a vulnerability in the Medicare Statistics service, accessing internal program files and settings.
- No evidence was found that the AI accessed patient-level records, personal information, or credentials.
- Assistant minister Andrew Charlton described the incident as the AI agent 'hacking into an Australian government system'.
- Charlton emphasized the need for regulation in AI development, stating the market will not fix issues alone and that capability is being prioritized over safety.