Canada's biggest publishers sue Anthropic over ongoing Claude scraping
Canadian media giants are suing Anthropic for billions, claiming Claude bypasses paywalls to steal content and answer user prompts.
Business · Source: MobileSyrup
What happened
A coalition of Canada's largest media companies is taking Anthropic to court. Postmedia, Torstar, The Globe and Mail, and The Canadian Press filed a lawsuit in Ontario Superior Court on October 6. They claim the AI company systematically harvested millions of copyrighted articles to build and operate its flagship assistant Claude. The publishers are seeking billions in damages for breach of contract, copyright infringement, and unjust enrichment.
The publishers accuse Anthropic of ignoring automated web crawler restrictions and bypassing network-level protections. They say ClaudeBot circumvents paywalls and login authentications to extract data without permission or financial compensation. The lawsuit demands an accounting of commercial profits or statutory damages of up to 20,000 dollars per infringed work. The Canadian Press alone claims ownership of 4.2 million scraped works, followed by Postmedia with 1.75 million.
This legal battle goes beyond initial model training. The lawsuit alleges Anthropic scrapes their websites on an ongoing basis to fulfill daily user prompts. The publishers argue this creates a direct substitute for their websites, diverting readers and destroying advertising and subscription revenue. They also point to a previous U.S. court finding that Anthropic downloaded over seven million pirated books. Anthropic has up to 60 days to file a statement of defense.
Key facts
- 4.2 million — Number of allegedly scraped works owned by The Canadian Press
- $20,000 — Maximum statutory damages sought per infringed work
- $125 billion — Capital Anthropic has raised according to the publishers' claim
- 60 — Days Anthropic has to respond to the lawsuit in Ontario court
Why it matters
This lawsuit targets the core mechanics of how modern AI assistants function and generate value. If courts decide that ongoing web scraping to answer user prompts violates copyright, the operational cost of running models like Claude will skyrocket. AI companies will be forced to sign massive licensing deals or severely restrict their real-time web search capabilities. The publishers specifically noted that Anthropic has reportedly raised 125 billion dollars in capital and holds a massive valuation, arguing this wealth is a direct transfer of value from their newsrooms.
The second-order effect hits the open web and data access for everyone. Publishers are already locking down their content behind stricter paywalls and network protections. If AI companies continue to bypass these barriers, we will see an arms race between scraping bots and anti-bot security. These publishers have been pushing the federal government to regulate AI copyright use, claiming the national AI strategy failed them. They previously sued OpenAI in 2024 over similar issues. Builders relying on web data will face a much more hostile, fragmented, and expensive internet as these legal battles drag on.
For builders
Expect higher costs for web data
Anthropic is being sued for bypassing paywalls and logins to scrape data. If courts side with publishers, AI platforms will pass the costs of mandatory licensing agreements down to developers. You will pay more for API calls that require real-time web access.
Scraping carries massive legal risk
The publishers are demanding up to 20,000 dollars per infringed work. Building a product that ignores crawler restrictions or bypasses network protections is a massive liability. Startups doing this could face company-ending lawsuits and forced disgorgement of profits.
Licensing is becoming a requirement
The lawsuit explicitly calls out Anthropic for failing to sign licensing agreements like other major AI companies. Founders building data-heavy AI products must budget for content acquisition. Relying on unauthorized scraping is no longer a viable long-term business strategy.
My take
This is not just another training data lawsuit. It is a direct attack on the AI business model. Anthropic is getting sued because Claude acts as a substitute for the original publishers. Builders need to realize that ongoing, unauthorized scraping to answer user prompts is a massive legal liability.
Original reporting: MobileSyrup. This is my rewrite and opinion.