Anthropic Backs 'Opt-Out' Rule for Australian AI Training as ABC Warns of News 'Cannibalisation'
In one line: Anthropic is reportedly pushing an opt-out approach in Australia's AI copyright debate, while public broadcaster ABC resists it as a threat that would cannibalise the news industry.
Key points
- Anthropic has reportedly argued, in Australia's policy review, for an opt-out model under which AI firms could train on content unless the rightsholder explicitly refuses.
- The public broadcaster ABC opposed the stance, warning it would lead to the "cannibalisation" of news content.
- The core dispute is which should be the default — prior consent (opt-in) or after-the-fact refusal (opt-out) — and therefore whether the burden falls on creators or AI companies.
Why it matters
If opt-out becomes the default, publishers would have to notify every AI company individually to refuse, effectively making training use the norm. The debate mirrors fights in the EU and UK, and Australia's decision could set a precedent for how news media control and are compensated for their content.