Publishers now have an official standard to tell AI crawlers exactly what they can and can't do with content. The Really Simple Licensing (RSL) 1.0 specification, released December 10, extends the familiar robots.txt file with machine-readable licensing terms, payment models, and granular usage controls.
RSL launched in September 2025 with backing from Reddit, Yahoo, and Medium. The 1.0 release solidifies the spec and adds new capabilities. Publishers can now permit search engines to index their content while blocking AI search applications from using it in generated answers. The spec introduces three new usage categories ("ai-all," "ai-input," "ai-index") to make this distinction possible.
The real teeth come from infrastructure. Cloudflare, Akamai, and Fastly have endorsed RSL 1.0, meaning CDN providers can enforce licensing terms at the network level. Cloudflare already processes over a billion HTTP 402 ("Payment Required") responses daily. Akamai committed to helping publishers uphold the controls they specify. Over 1,500 media organizations now support the standard, including The Associated Press, The Guardian, USA Today, BuzzFeed, Stack Overflow, and Vox Media.
One catch: no major AI company has committed to honoring RSL 1.0. OpenAI, Google, Anthropic, and Meta have made no public statements about reading or respecting the spec. Enforcement currently depends on CDN-level blocking rather than voluntary compliance from model builders.
The Bottom Line: RSL 1.0 creates the licensing vocabulary publishers need, but whether AI companies will pay attention remains an open question.
QUICK FACTS
- Released: December 10, 2025
- Supporters: 1,500+ media organizations
- CDN backers: Cloudflare, Akamai, Fastly
- New usage categories: ai-all, ai-input, ai-index
- Technical lead: Eckart Walther (RSS co-creator)
