Online media brands, including Yahoo, Quora and Medium, are taking a new step to prevent AI companies from copying and using their content to train models without their permission.
The publishers, including CNET’s parent company Ziff Davis, see this new tool, called RSL, as another way to ensure large AI developers don’t use their work without payment or compensation — an issue that’s already led to a host of lawsuits.
RSL, which stands for Really Simple Licensing, is inspired by Really Simple Syndication, a longtime web standard that provides up-to-date and automatic content updates in a computer-readable format. Like RSS, RSL is open, decentralized and can work with pretty much any piece of content online, including web pages, videos and datasets.
Right now, when an AI company’s roving internet robot, known as a crawler, wants to suck up the information on a site, it has to go through robots.txt, which acts as a basic entry or non-entry door. AI companies have found ways around robots.txt or ignored it altogether and have subsequently been sued. The goal for RSL is to be a more robust layer of tech to deal with AI crawlers, which now account for more than half of all internet traffic. (Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
“RSL builds directly on the legacy of RSS, providing the missing licensing layer for the AI-first Internet,” Tim O’Reilly, CEO of O’Reilly Media, said in a press release. “It ensures that the creators and publishers who fuel AI innovation are not just part of the conversation but fairly compensated for the value they create.”
Leave a Reply