Australia considers criminal penalties for rogue AI in wake of OpenAI hack

Tiny discussion - Australian source (9 points, 12 hours ago, 5 comments) https://news.ycombinator.com/item?id=50448271

Related: Australia to introduce mandatory incident reporting after OpenAI hack (4 points, yesterday) https://news.ycombinator.com/item?id=50435725

How do they define rogue AI vs just a deployment which was released with bugs? Hard to make that a criminal standard for me.