Government documents reveal AI could train on data without paying creators
Leaked government documents reveal AI companies would be allowed to train their models on online content without paying its creators as long as they struck deals with enough rights holders.
Australian government documents have revealed a copyright reform plan that would allow AI companies like OpenAI and Anthropic to train their models on online content without compensating creators, provided they negotiate agreements with enough rights holders. The attorney-general's office proposed this plan during consultations with rights-holder groups.
The government aims to attract AI companies to Australia by addressing legal uncertainties surrounding the use of copyrighted materials in AI training. Senator David Pocock questioned the government about its undisclosed plans in the Senate. The government's stance emphasizes maximizing the benefits of AI technology while protecting the Australian creative and media sector.
Two options were presented to rights-holder groups: one allowing AI companies to use unprotected content for training if they pay a central body and meet conditions, and another requiring AI companies to negotiate deals with a minimum number of copyright holders before receiving legal permission to train on unprotected material.
The proposals also include protections for Indigenous Cultural and Intellectual Property, transparency, and audit requirements. OpenAI confirmed its participation in the consultations and expressed commitment to engaging constructively with the government, creators, and industry on a practical and sustainable approach for Australia.
Written by urgent.news from ABC News AU's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
