Pay Up, AI Bot: That’s the Message From a Key Company in How the Internet Works

AI companies can find that Internet infrastructure provider CloudFlare is this week to get large language models are difficult to access AI would block data scans as standard.

The creators of content and this content is the latest front of the ongoing battle between AI developers using the AI ​​developers Generative EU models. In court, the creators of the authors and content, copyrighted content in court shall sue the basic AI companies for compensation. .

While compensation for the information used to prepare models in the past, CloudFlare’s movement marks a new defense size against future efforts to prepare models.

But this is not only related to scanning: CloudFlare says it wants create a market AI receives a permission of the AI ​​developer where the EI can pay for the procurement and crawl of a site where this information is paid.

“This content is fuel that strengthens AI engines, and therefore is fairly compensated for the creators of content”, “Coughtare CEO Matthew Prince A Blog Post.

Why websites want to block AI reptiles

Severes – Bots that visit the information and copying data from a website – a vital part of the connected Internet. How do search engines like Google are how different websites are and how you can serve the latest information from places like CNET.

AI reptiles create different problems for websites. For one, they can be aggressive, can be created Unstable level of traffic for small sites. They offer a small reward for their stutter: If Google is dragged into a site for search engines, it will send traffic to those sites, including search results. Creeping for training information, people stop visiting the site and only do not trust the AI ​​model, there may not be additional traffic or less.

Read more: ESSENTIALS: According to our experts, Gen AI 29 ways to work for you

Therefore, such as Pinterest, Reddit and several major publishing companies, the basic websites (including CNET Davis, the owner of CNET) pleases CloudFare news in statements.

“The whole ecosystem of the Creators, platforms, web users and reptiles will be better and more transparent and Cloudfare’s efforts are a step in the right direction for everyone,” said Reddit Cee Huffman said.

CloudFlare’s announcement asked, Openai, ChatGPT model is designed to help users connect the content of the content on the Internet and access to search engines similar to search engines. The company also said that the clouds use a model proposed to ensure how to manifest their behavior, known as the Publishers of the clouds. Openai said the robots.txt is already working and Cloudfare’s changes are unnecessary.

Training information Tug-of-war

AI models require a ton of information for the train. These are able to answer the questions in detail and do a decent (flawless) job to provide extensive information. These models are fed incredible amounts of information and are in touch between words and concepts based on what you see in this training information.

The issue was how it developed this information. There is already Dozens of courts Among content creators and AI companies. He saw two great judgments last week.

In somethingThe Federal Judge, Anthropiki, followed by the law when used with copyrighted books to train the model clod – through a concept called fair use. At the same time, the ruling company said the company’s creation of a permanent library of books and was new to this pirate allegations.

He ruled a judge in a separate In favor of the metage The company and 13 in the dispute between a group of author. However, in this case, in this case, in this case, in this case, in this case, the judge will go to the integers and other EU companies, “these bidders made wrong evidence and could not develop a record.”

The idea of ​​charging to visit a site to the shores is not completely new. Other companies, like TollSuggest websites that allow you to send AI companies to creep. AI control, President of Privacy and Media Products in Tollbit, said the environment around this technology is still developing. “We think we are very early for a content market and we are just starting to practice here,” he said. “We are pleased to see many different models flourish.”

CNET’s Imad Khan contributed to this report.