“Really Simple Licensing” – The Path to a Fair Digital Content Economy?

Over 40 copyright lawsuits have been filed against AI companies in the U.S. since 2022. These cases seek compensation for the plaintiffs for the unauthorized use of their data by AI models. The need to find a fundamental solution to this issue is therefore clear.

How does Really Simple Licensing work?

Really Simple Licensing is an open standard that provides the licensing terms of a web resource in a machine-readable format for the bots of AI companies. From a TXT file (named rsl.txt), the bots learn the names of the creators or rights holders of the content, as well as the terms for data usage and compensation. The RSL Standard 1.0 has been officially in effect since December 2025.

RSL relies on usage-based licensing fees, ensuring that creators and rights holders are compensated whenever their data is used by AI.  For AI companies, this offers the ability to calculate licensing costs based on frequency of use. Even non-public content, such as articles and books behind a paywall, can be included in the RSL system and thus utilized by AI companies.

What does Really Simple Licensing offer creators?

In a “Creator Bill of Rights for the AI Era,” RSL sets forth three key commitments to creators and rights holders of creative content—whether it be text, images, videos, audio, datasets, or other formats.

1. Control over usage. Content creators and rights holders should be able to control who uses their content and for what purposes.
2. Representation of interests vis-à-vis the AI industry. All rights holders should be able to negotiate compensation collectively with AI companies.
3. Compensation for every use. RSL 1.0 offers content creators and rights holders various compensation options. If content is used by an AI to produce a task result, this can be compensated through attribution and linking. In the case of monetary compensation, payment can be made either for the collection of content by an AI (pay-per-crawl) or for the use of the content in a task result (pay-per-inference). Free access and a subscription model are also possible.

What does Really Simple Licensing offer the AI industry?

Creativity should, therefore, be fairly compensated. For AI companies, RSL primarily promises increased efficiency. Automating the licensing process is intended to avoid cumbersome, ad-hoc technical solutions. Standardized licensing terms are also designed to eliminate the need for time-consuming negotiations with individual rights holders.

Access to data is to be made easier overall. This is to be achieved, on the one hand, through access to non-public content that is protected from unauthorized access by encryption. But efficiency is also gained through the legal safeguarding of data use, as this includes transparent documentation of the copyright status of any data used.

What is Really Simple Licensing’s stance on CC Signals?

The role of the Creative Commons (CC) organization in the context of the RSL initiative is particularly interesting. CC’s current licensing model permits AI training using CC-licensed content; the only restriction applies to commercial use, as governed by the NC license module. However, Creative Commons introduced the new concept of “CC Signals” last year.

Our author Fabian Rack describes CC Signals as a concept currently under development, through which CC is responding to the “breach of the social contract of openness.” The goal of this CC approach is to address the “distortions caused by AI without resorting to stricter copyright laws.” Instead, in addition to appropriate attribution of rights holders (“appropriate credit”) or open use (“Open”), CC Signals also provide for two elements that mandate compensation in the form of monetary or in-kind payments: “Direct Contribution” and “Ecosystem Contribution.”

RSL 1.0 is now the first application to be implemented as part of the CC Signals initiative. If the RSL standard succeeds in establishing itself, CC Signals will thus be brought from the conceptual level to the application level.

“Reciprocity” in Really Simple Licensing

Fabian Rack points out that CC Signals aims to establish “a new give-and-take between content creators and AI providers,” and that CC refers to this as “reciprocity.” In the course of the preliminary negotiations between Creative Commons and the RSL Collective, this “reciprocity” has been incorporated as a separate component for RSL’s usage compensation: “A monetary or in-kind donation made in good faith that supports the development or maintenance of the assets or the broader content ecosystem.” Thus, the “Direct Contribution” and “Ecosystem Contribution” from CC Signals are implemented in the RSL standard.

From CC’s perspective, the compensation contribution could, for example, take the form of a donation to a nonprofit organization that manages the relevant dataset. The “broader ecosystem” surrounding a dataset could also benefit financially. Other possible approaches include open licensing of an AI model trained on the data, or making an AI-“modified dataset” available to the data steward of the original dataset. Creative Commons emphasizes that “entirely different models that we cannot yet imagine” may also be added in the future.

What are the prospects for Really Simple Licensing?

The RSL Standard 1.0 has only been active for a few weeks. A number of major companies in the internet industry are already publicly supporting RSL. These include infrastructure companies such as Akamai and Cloudflare, and social media services like Reddit and Stack Overflow. Support from the media industry ranges from A for Associated Press to Z for ZDNET.

The rapidly growing support from content producers is not very surprising. After all, the new standard promises fair compensation for creative work. As is well known, the AI industry is expected to pay these royalties. However, the industry has so far remained tight-lipped about RSL. “Time will tell” is one comment on the matter.

Jens Crueger ist Redakteur bei iRights.info. Er ist Historiker mit Schwerpunkt Digital History, Geschichte des Web und digitale Archivierung. Er promoviert in digitaler Medizinethik an der Fakultät für Gesundheitswissenschaften (FGW) Brandenburg.