On Wednesday, Cloudflare’s online infrastructure provider announced a new feature called “Amnesty International Maze“This aims to combat unauthorized artificial intelligence data by providing fake content of artificial intelligence with robots. The tool will try to thwart companies that crawl on web sites without permission to collect training data for large language models that operate AI assistants such as ChatGPT.
The Cloudflare, founded in 2009, is likely to be known as a company He provides Infrastructure and security services for web sites, especially protection against Distributor for the service refusal (DDOS) attacks and other malicious traffic.
Instead of just banning robots, the new Cloudflare system is dominated by “maze” of real but unrelated appearance pages, which istes computing resources in the crawling. This approach is a noticeable shift from the standard strategy and standard defense used by most site protection services. Cloudflare says that the prohibition of robots is sometimes reversed because it alerts the creeping operators that were discovered.
“When we discover the unauthorized crawl, instead of preventing the request, we will link with a series of pages created from artificial intelligence that persuade enough to lure a creeping to cross them.” “But despite the real appearance, this content is not in fact the content of the site we protect, and therefore the creeping is waste time and resources.”
The company says that the content that is presented in robot programs is not relevant to the site crawling, but it is obtained carefully or generated by using real scientific facts – such as neutral information about biology, physics or mathematics – to avoid spreading misleading information (whether this approach actually prevents wrong information). Cloudflare creates this content using it Artificial intelligence workers The service, a commercial platform that manages artificial intelligence tasks.
Cloudflare designed the trap pages and links to remain invisible and unknown to ordinary visitors, so people who browse the web do not collide by chance.
Honips is smarter
AI Labyrint works as Cloudflare calls it “a generation of the next generation”. Traditional honeymaker are invisible links that human visitors cannot see, but this may follow robots that are analyzed HTML. But Cloudflare says that modern robots have become brilliant in discovering these simple traps, which requires more sophisticated deception. Wrong links contain appropriate metasta directions to prevent the search engine indexing while maintaining attractive data robots.