The Australian Financial Review
Details
- Date Published
- 17 Sept 2026
- Priority Score
- 3
- Australian
- Yes
- Created
- 18 Sept 2026, 12:00 am
Authors (2)
- Sam Buckingham-JonesENRICHED
- Zoe SamiosNEW
Description
A proposal that would give AI giants free rein to train their models on copyright-protected work appears to have been abandoned by the attorney-general.
Summary
The Australian government has reportedly abandoned a proposal that would have granted AI companies like OpenAI and Anthropic broad rights to train models on copyright-protected Australian content. This policy pivot follows intense pushback from domestic media and creative industries regarding compensation for intellectual property used to develop frontier AI capabilities. While primarily focused on copyright and economic incentives, the decision reflects broader governance tensions in managing the data pipelines necessary for advanced AI development. This development is a key indicator of Australia's evolving regulatory stance on how large-scale model training interacts with national creative and economic frameworks.