Reddit, yahoo, medium and more adopt a new standard of licensing for and scraping

With Web publishers in crisisThe new open standard allows them to set the basic rules for AI scrapers. (Or, at least will try.) New really easy licensing (RSL) Standard creates Conditions that participants expect companies to adhere to. Although implementation is an open question, he cannot harm some heavy blow to him. Among other things, the list includes reddit, Yahoo, Middle and People Inc.

RSL adds notes for licensing on robots.txt protocol, simple file that provides web stick instructions. Supported licensing options include free, attribution, subscription, payment per declinion and payment per conclusion. (The latter means that AI companies pay only publishers when the content is used to generate answers.)

Starting together with the standard is a new management of non-profit, RSL collectively. This is displayed as an equivalent of non-profit products such as ASCAP and BMI, which manage the kings of the music industry. The new group says that his standard can “establish fair market prices and strengthen the negotiating plant for all publishers”.

The brands of participants include the abundance of old tuition fees. Reddit, People Inc., Yahoo, Internet brands, Ziff Davis, Vikihov, O’Reilly Media, Medium, Daily Beasts, Miso.ai, Raptive, Rapila and Divorce Media are on board. Former Ask.com General Manager Doug Leeds and RSS Co-Creator Eckart Walther leads a group.

“RSL Standard gives publishers and platforms a clear, scalable way to determine the notions of licensing in AI era,” Reddit CEO Steve Huffman wrote in a statement to the public. “RSL collective offers the way to do it together. Reddit also supports as important steps to protect open web and communities that make it work.” (It is worth noting Reddit has agreed offers with Openai and Google.)

It is unclear whether the companies and honor the standard. After all, they knew that simply Ignore robots.txt instructions. But the group believes that his conditions will be legally executive.

In an interview with Ars TechnicaLeeds index to anthropic recently 1.5 billion dollars of settlementsSuggesting “It’s a real money” for AI companies that don’t train “legitimate”. (However, this settlement is in the air after a The judge rejected him.) Leeds told Edge In order for the standard collective nature could also help spread legal costs, making challenges by feasible violations.

As for technical implementation, RSL standard cannot block bots on its own. Therefore, the Group is a fast partner with Cloud Companium, which can act as a kind of goalkeeper. (Maybe Cloudflare, which is recently launched the payment system per creepIt could also eventually play part.) Leeds said quickly that I could serve as “throw at the door at the club.”

Lieds suggested Ars That there are also incentives for AI companies. Financially, it could be simpler for them than Indignation of individual licensing agreements. This could prevent a problem in AI content: using more response sources to avoid using too much than any. If the content is legally licensed, and the application can easily use the best source that provides the user with higher quality answers and minimizes the risk of hallucinations.

He also complained from AI companies that there are no effective funds for licensing content on the site. “We listened to them and what we heard say it was … We need a new protocol,” Leeds said Ars Technica. “With the Rsl Standard, AI Firms Get a” Scalable Way to Get All The Content “They Want, The Best Content That Their Models Actually Reference. If They’re using It, They Pay For It, And If They’ve Not using It, They Don’t Pay For It. “