National framework for AI assurance in Australian government: Guidance when building with AWS AI/ML solutions

This article provides guidance to Australian government agencies on implementing AI/ML solutions responsibly and in alignment with the National Framework for the Assurance of Artificial Intelligence in Government. It covers how AWS tools and services can support agencies in adhering to Australia's AI Ethics Principles.

Specifically, the article covers:

• An overview of the National Framework and its eight AI ethics principles
• How AWS supports each principle:
  • Human, societal and environmental wellbeing (e.g. carbon footprint tracking, energy-efficient processors)
  • Human-centred values (e.g. diverse teams, external resources)
  • Fairness (e.g. bias detection with SageMaker Clarify)
  • Privacy and security (e.g. privacy impact assessments, security controls)
  • Reliability and safety (e.g. model evaluation, watermarking)
  • Transparency and explainability (e.g. SageMaker Clarify, Model Cards)
  • Contestability (e.g. feedback loops, appeals processes)
  • Accountability (e.g. defining roles, training resources)
• Conclusion: AWS is committed to helping Australian agencies build responsible AI solutions aligned with the ethical principles.