Like many other countries, controversy currently surrounds generative AI and copyright in Australia, driven by debates over how large language models (LLMs) ingest and utilize local creative works.
Affected Copyright Works
The works primarily impacted include books, journalism, musical compositions, sound recordings, artistic works, and screen content created by Australian authors, journalists, musicians, and artists.
AI Misuse and Core Issues
The primary "misuse" involves unauthorized scraping and ingestion by LLMs, where tech companies use massive datasets of potentially copyrighted Australian text, art, and music to train LLMs via text and data mining (TDM) without permission or compensation. Key stakeholders include bodies like the Australian Publishers Association (APA), the Australian music industry, and various arts, media, and tech sector participants collaborating through the Copyright and Artificial Intelligence Reference Group (CAIRG), alongside the Federal Government