Anthropic Signs First News Licensing Deals, Reversing Previous Stance
Anthropic has signed its first news content licensing agreements with major publishers, marking a significant reversal in the AI company’s approach to using journalism for training its AI models. The deals with News Corp and Axel Springer will allow Anthropic to use news content to train Claude and power AI-generated responses with cited news material.
The move positions Anthropic alongside competitors like OpenAI and Google, which have already struck similar partnerships with publishers worldwide.
A Major Reversal
Anthropic had previously resisted paying for news content. The company argued that training AI models on publicly available text fell under the fair use doctrine.
This shift signals that Anthropic now recognizes the need to compensate publishers for their work, at least in certain contexts.
The company faced mounting legal pressure from copyright holders. Several publishers had sued AI companies over the unauthorized use of their content in training datasets.
What the Deals Include
The licensing agreements cover:
- News Corp publications: The Wall Street Journal, Barron’s, and other outlets under the News Corp umbrella will be available to Claude.
- Axel Springer properties: Politico, Business Insider, and Bild content will also be licensed for use.
- Training and product features: The content will be used both to train AI models and to provide cited answers in Claude.
The financial terms of the deals were not disclosed. However, industry reports suggest these agreements are substantial, given the scale of the publishers involved.
Why This Matters for the AI Industry
The deals represent a significant moment in the ongoing tension between AI companies and news publishers.
- Legal pressure is working: Publishers have been aggressive in pursuing legal action against AI companies for copyright infringement.
- Licensing is becoming the norm: OpenAI, Google, and now Anthropic have all moved toward licensing agreements rather than fighting in court.
- The cost of AI training is rising: As more publishers demand compensation, the cost of building and maintaining AI models will increase.
The Broader Copyright Battle
Anthropic’s reversal comes amid a wave of lawsuits and negotiations across the AI industry. News organizations have argued that AI companies are profiting from their work without compensation.
Some publishers have chosen to negotiate licensing deals. Others have taken a more adversarial approach, filing lawsuits in multiple jurisdictions.
The outcome of these battles will shape how AI companies access and use copyrighted material in the future.
What This Means for Claude Users
For users of Claude, the deals mean access to more current and authoritative news content. When asking Claude about recent events, the AI may now be able to draw on licensed news sources and provide citations.
This could improve the accuracy and reliability of Claude’s responses on news-related topics. It also reduces the risk of the AI reproducing copyrighted text without permission.
The Bottom Line
Anthropic’s decision to pay for news content represents a pragmatic shift. Rather than continuing to argue that fair use covers their training practices, the company has chosen to negotiate with publishers.
This is likely the beginning of a broader trend. As legal pressure mounts, more AI companies will probably follow suit and sign licensing deals with news organizations.
For publishers, these deals provide a new revenue stream. For AI companies, they provide legal cover and access to high-quality content. For users, they mean better AI products.
The era of AI companies using news content for free appears to be ending.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.