Response from Anthropic’s Claude.ai:
Usage-based deals represent a promising middle ground between the current landscape of content licensing and the publishers’ concerns about AI training on their content. Here are several models LLMs could offer publishers:
Revenue-Sharing Models
- Pay publishers a percentage of revenue generated when their content is used to answer user queries
- Implement tracking systems that identify when specific articles, datasets, or publications contribute to responses
- Create tiered revenue sharing based on content quality, exclusivity, or user engagement metrics
Pay-Per-Citation Systems
- Compensate publishers each time their content is directly cited or referenced in AI responses
- Offer premium rates for exclusive content that gives the AI a competitive advantage
- Scale payments based on the prominence of the citation (primary source vs. supporting reference)
Content Licensing Tiers
- Basic tier: Standard web crawling with minimal compensation
- Premium tier: Real-time access to breaking news, exclusive content, or high-value databases
- Enterprise tier: Custom content partnerships with guaranteed attribution and higher revenue shares
Performance-Based Arrangements
- Pay based on user satisfaction metrics when publisher content is used
- Bonus payments for content that drives user engagement or retention
- Quality premiums for fact-checked, authoritative sources
Subscription Revenue Sharing
- Share a portion of AI service subscription fees with publishers whose content is frequently accessed
- Create publisher pools where compensation is distributed based on content utilization analytics
Attribution-Linked Advertising
- Drive traffic back to publisher sites through prominent source linking
- Share advertising revenue from users who click through to original sources
- Co-branded content experiences that benefit both AI platforms and publishers
These models would require sophisticated content tracking, transparent analytics, and new technological infrastructure to fairly measure and compensate content usage. The key is creating win-win scenarios where publishers see tangible value from AI platforms using their content.