ChatGPT broke last Tuesday — bots are to blame

ChatGPT broke last Tuesday. This is the biggest failure since 2019, wrote the CEO of Cloudflare. The reason is the bot management system. It must control how automated robots can crawl certain websites.

Cloudflare told xrust last year that about 20% of the Internet flows through its network, which should distribute the load, keeping websites running smoothly in the face of traffic spikes and DDoS attacks.

Yesterday, however, there was an outage disabled many of them for several hours, disabling everything from X to ChatGPT and the famous Downdetector crash tracker. The situation was reminiscent of recent outages caused by problems with Microsoft Azure and Amazon Web Services.

The problem is caused by Cloudflare's bot management. The system is designed to help solve issues such as data collection by search robots for training generative AI. A system using generative AI to create an “AI Labyrinth” was recently announced. This is a new approach to reducing risks. They are caused by the use of AI-generated content to slow down, confuse and waste AI robots and other bots that do not adhere to «no crawl» directives.

Yesterday's problems were caused by changes to the database permissions system, not generative AI technology, not DNS, and not what Cloudflare initially suspected — a cyberattack or malicious activity like “hyperscale DDoS attack.”

In general, they got confused when regulating.

Xrust ChatGPT broke last Tuesday — bots are to blame