KEY POINTS
- Australia will consider a new criminal offense for AI companies that knowingly release harmful models.
- A rapid review made 18 recommendations, including mandatory reporting of security incidents and misaligned model behavior.
- The government also wants to attract AI datacenter investment, which may limit how hard it pushes on penalties.
The Australian government will consider creating a criminal offense aimed at AI companies that knowingly release harmful models, The Register reported. The idea appeared in the results of a review that Canberra published a day before the outlet's 11 November report.
