Published on 01/09/2025 - SEPM press release
The General Information Press Alliance and the Magazine Press Publishers Union announce the launch of a coordinated action aimed at obtaining the removal of their members' content from the public databases Common Crawl, C4 and Oscar, which are massively used by generative artificial intelligence services for training.
An ecosystem for laundering publishers' content
This initiative responds to an alarming observation: generative AI providers are massively sourcing press content via so-called "public" databases which reproduce and distribute millions of articles protected by copyright and related rights without authorization or implementation of any access restrictions whatsoever.
In fact, these datasets constitute a veritable ecosystem for laundering unauthorized uses, allowing generative AI service providers to circumvent the law by using supposedly open-access data.
A coordinated strategy to restore a fair balance in accordance with the law
The action undertaken by the Alliance and the SEPM has three objectives:
A defense of the economic model of professional information
This action carries a fundamental principle: the production of professional information requires investments that must be fairly remunerated. The publishers of the Alliance and the SEPM, who employ 57% of French journalists, are thus defending the economic viability of professional and quality journalism, which constitutes an essential guarantee for a democratic society.
The initiative is part of the overall strategy of the Alliance and the SEPM aimed at enforcing the intellectual property rights of publishers.