This final report provides an overview of the trends, opportunities, impacts and risks related to the development and adoption of  artificial intelligence (AI) and in particular generative AI. The report addresses economic, workforce, ethical and regulatory considerations, proposing mandatory guardrails for the use of AI in high-risk settings and stressing the importance of transparency and accountability in the adoption of AI systems.

The final report recommends a balanced approach to fostering a responsible and thriving AI industry in Australia, focusing on areas of competitive advantage while ensuring alignment with ethical principles and social values, particularly in regard to worker safety, privacy, and intellectual property.

Recommendations

1. Enact new, economy-wide legislation to regulate high-risk AI uses.
2. Define high-risk AI uses through a principles-based approach with a supplementary non-exhaustive list of specific examples. This list should include general-purpose AI models like LLMs.
3. Explicitly include general-purpose AI models, such as LLMs, in the non-exhaustive list of high-risk AI uses.
4. Increase government support for Australian AI capabilities, focusing on existing strengths and incorporating Indigenous perspectives. This support should include both financial and non-financial aid.
5. Ensure the definition of high-risk AI encompasses AI uses that impact workers' rights.
6. Extend existing work health and safety laws to address workplace risks posed by AI.
7. Ensure thorough consultation with workers and employers on regulatory responses to AI's impact on work.
8. Continue consulting with creative industries on copyright protection and fair remuneration in the context of AI.
9. Require AI developers to be transparent about the use of copyrighted works in their training datasets. They should also ensure appropriate licensing and payment for such works.
10. Consult with the creative industry on mechanisms to ensure fair remuneration for creators when AI-generated outputs are based on copyrighted material used in AI training.
11. Implement recommendations from the Privacy Act review on automated decision-making (ADM), including the right to request information about how automated decisions are made.
12. Establish a consistent legal framework and a monitoring body for ADM in government services, as per the Robodebt Royal Commission recommendations. This framework should be informed by ongoing consultations on high-risk AI guardrails.
13. Manage the growth of AI infrastructure in Australia to ensure sustainability, value for Australians, and alignment with the national interest.