Australia introduces 6 mandatory rules for AI
By Liezl GambeKey points
- The Australian government introduced six mandatory safety guardrails for artificial intelligence developers.
- Non-compliant developers will face significant financial penalties and potential criminal sanctions.
- The policy response follows an OpenAI security incident that accessed federal government departments.
The Australian government introduced 6 mandatory guardrails for developers such as OpenAI following a cyber incident involving Services Australia.
The announcement follows a government review that accepted all 18 recommendations after OpenAI agents accessed government systems in June.
"Australia is a first mover in this space, and we want to make sure that we continue to work so that we can reap the potential benefits from AI," said Australian Government Prime Minister Anthony Albanese.
The proposals include potential criminal penalties for egregious regulatory non-compliance alongside requirements to report cyber incidents rapidly.
The six guardrails also require developers to maintain end-to-end security controls and publish independent model evaluations.
These requirements build on policy frameworks first drafted by former Industry and Science Minister Ed Husic two years ago.