Australia: Australia proposals paper for introducing mandatory guardrails for AI in high-risk settings (10 guardrails, testing and accountability)

1. 1. Delivering regulatory clarity and certainty
2. 2. Supporting and promoting best-practice governance
3. 3. Supporting AI capability
4. 4. Government as an exemplar
5. 5. International engagement
6. 6. Should mandatory guardrails apply to all GPAI models?
7. 1. Connectedness – Polarisation
8. 2. Transparency – Opacity
9. 3. Decentralisation – Consolidated control
10. 4. Democratising voices – Narrowing voices
11. 5. Truth and deliberation – Deception
12. 6. Public good – Private gain
13. 7. Information engagement – Information transmission
14. 2. Establish and implement a risk management process to identify and mitigate risks
15. 3. Protect AI systems, and implement data governance measures to manage data quality and
16. 4. Test AI models and systems to evaluate model performance and monitor the system once
17. 7. Establish processes for people impacted by AI systems to challenge use or outcomes
18. 9. Keep and maintain records to allow third parties to assess compliance with guardrails
19. 1. Potential gaps in regulatory coverage
20. 2. Consistency within but not across sectors
21. 5. International engagement on AI governance
22. 7. Establish processes for people impacted by AI Deployers are to establish internal

The outline comes from headings detected in the text, so it might miss some or include extras.