With the announcement from the Albanese government that it will not introduce a text and data mining exemption into Australian copyright law, we’ve seen the temporary end of the first of many proxy battles in what is sure to be a long war of attrition over artificial intelligence.
At the heart of the fight is training data. Every major large language model, whether it’s OpenAI’s GPT, Anthropic’s Claude or Google’s Gemini, is built on vast libraries of copyrighted work. Rights holders are compensated inconsistently, if at all.
The result has been significant litigation: Anthropic agreed last month to pay USD1.5 billion ($2.2 billion) to settle a class action from authors who said their books were trained on without permission.
Down Under, it became something of a trial by fire for the Tech Council of Australia under the leadership of chair Scott Farquhar, who was largely responsible for forcing the matter of tech-friendly copyright reform into the national discourse.
