Anthropic Urges Australia to Allow AI Training Opt-Out as ABC, SBS Warn of News Cannibalisation
Updated
Updated · The Guardian · Oct 1
Anthropic Urges Australia to Allow AI Training Opt-Out as ABC, SBS Warn of News Cannibalisation
1 articles · Updated · The Guardian · Oct 1
Summary
Anthropic told an Australian parliamentary inquiry to grant “conditional approval” for AI training on copyrighted works through an opt-out system, after the Albanese government ruled out a broad text-and-data-mining exemption.
Robots.txt “do not crawl” signals are Anthropic’s proposed safeguard, but SBS said such tools are often ignored or bypassed and unfairly shift the burden of protection onto rights holders.
ABC and SBS urged stricter rules requiring AI firms to follow the same copyright, defamation and privacy standards as media outlets, and to pay publishers through Australia’s news bargaining incentive.
Anthropic argued AI could lift productivity and justify billions in Australian datacentre investment, even as it acknowledged risks to national security, critical infrastructure and public safety.
Sydney hearings start next week with Anthropic and OpenAI executives due to appear, keeping copyright, compensation and AI oversight at the center of Australia’s policy debate.