Anthropic Proposes Opt-Out Training Model as Australian Broadcasters Warn of News Cannibalisation
Anthropic is pushing Australia to allow AI training via an opt-out model, but public broadcasters ABC and SBS warn the move threatens the survival of local journalism.
Anthropic has proposed a "conditional approval" framework to the Australian government that would allow tech companies to train their artificial intelligence models on copyrighted works using an opt-out system. The submission to a federal parliamentary inquiry comes as the country's public broadcasters, the ABC and SBS, raise alarm over the unregulated nature of AI development, warning that it could lead to the "cannibalisation" of local mainstream news.
The debate highlights a growing tension between frontier AI developers seeking vast datasets and international creative industries demanding compensation and protection. While Prime Minister Anthony Albanese's Labor government has already ruled out a broad text and data mining exemption for AI training, Anthropic is pushing for a compromise that would provide legal certainty while allowing publishers to choose whether their content can be accessed.
The "Opt-Out" Proposal and Publisher Backlash
In its parliamentary submission, Anthropic acknowledged that a blanket copyright exemption has failed to win widespread support. Instead, the company suggested a narrower approach where AI training is permitted unless content creators explicitly signal otherwise. Anthropic proposed using standard web protocols, such as "do not crawl" instructions in robots.txt files, to let copyright holders block AI bots from accessing their material.
However, Australian public broadcasters have strongly rejected this approach. SBS argued that an opt-out model is fundamentally flawed, stating that robots.txt mechanisms are "routinely ignored or bypassed." The broadcaster emphasized that forcing rightsholders to constantly monitor and update these blocks places an unfair burden on creators, when the responsibility should instead lie with AI companies to avoid using unauthorized material.
The ABC went further, describing the current landscape as "anti-competitive" because AI-based services remain unregulated while traditional media outlets face strict legal frameworks. In its submission, the ABC warned that this unequal treatment would accelerate the decline of the domestic news industry. Both the ABC and SBS have called on the government to extend the news bargaining incentive—which currently applies to platforms like Google and Meta—to AI developers, forcing them to negotiate commercial deals to support public interest journalism.
Economic Promises and Security Risks
Anthropic’s submission also outlined the broader stakes of AI development in Australia. The company pointed to the "light and shade" of the technology, suggesting it could drive productivity and solve major national challenges. To support this, Anthropic highlighted its plans to invest billions of dollars in Australia, including the construction of large new data centers, and urged the government to support "frontier model training."
At the same time, Anthropic warned of "serious implications" regarding Australia's national security, critical infrastructure resilience, and public safety. These issues will be discussed further at upcoming parliamentary inquiry hearings in Sydney, where executives from Anthropic and OpenAI are scheduled to appear. The hearings follow a recent controversy in June, where OpenAI admitted that one of its AI agents had accessed data from Services Australia.
What it means for developers
For software engineers and AI developers, the regulatory battle in Australia signals a shifting landscape for how foundational models are trained and deployed. If Australia implements strict copyright protections or forces AI companies into news bargaining frameworks, it could set a precedent for other jurisdictions, potentially limiting the availability of high-quality local data for training global models.
As compliance requirements grow more complex, developers may find it increasingly difficult to navigate the legalities of self-hosting or fine-tuning models on scraped web data. To mitigate these regulatory and infrastructure hurdles, many teams are turning to unified API services that handle model access and compliance on the backend. For example, developers can try top AI models cheaply through one API at https://apixoai.online, allowing them to experiment with models from Anthropic, OpenAI, and other leading providers without managing individual vendor agreements or worrying about regional scraping compliance.
Ultimately, the outcome of the Australian inquiry will help determine whether the future of AI training relies on open web scraping with opt-out mechanisms, or a highly regulated, licensed ecosystem where developers must pay for the data that powers their systems.
Source: Anthropic pushes for opt-out model for Australian content as ABC warns of ‘cannibalisation’ of news — The Guardian AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
