I reckon there's a high degree of probability whoever at OpenAI decided to notify Australia of its 'misalignment' attack by emailing the badnews@gov.au inbox had no idea the company has a government relations team who could have made a more graceful disclosure. That could be bad internal comms, or just a company growing too fast to even think about how to set rules for technical teams. This is not an excuse by the way - I have always said there's no such thing as a PR crisis. It's always a crisis caused by bad behaviour. In this case, OpenAI not monitoring its agents during tests. FFS, these people.