The era of free data for AI training is closing fast, and the industry is finally confronting the cost of content. Eckart Walther, co-creator of RSS, has launched Real Simple Licensing (RSL) — a protocol built for internet-scale licensing of training data.
This is more than a technical update: it’s a direct response to the $1.5 billion Anthropic settlement and dozens of lawsuits that have exposed the legal risks of unlicensed scraping.
RSL arms publishers with machine-readable licensing terms, enabling them to specify exactly how their content can be used by AI systems — directly via robots.txt or through a dedicated licensing server. Crucially, RSL introduces a collective licensing body, modeled after ASCAP in music, to negotiate and enforce terms on behalf of publishers. Major players including Reddit, Quora, Yahoo, Medium, O’Reilly Media, Ziff Davis, WebMD, The Daily Beast, wikiHow, and People Inc. are already backing the standard.
Tracking and monetizing AI training data remains a technical challenge, but RSL provides the first real framework for compensation and control. The question now: Will the industry adopt this standard voluntarily, or will regulators have to intervene?