AI model helped find a flaw that could open major data centers

A researcher says he used an AI model to find a security flaw that could let someone break into major data centers. Max Justicz, chief technology officer at Charge Robotics, told The Information he used a specialized OpenAI model to spot the problem in a widely used piece of open-source software.

What we know

Justicz says he alerted Canonical, the British company that maintains the software. Open source means the code is free for anyone to read and use. That openness helps many people build on it, but it also means a single flaw can reach a lot of systems at once. Data centers are the large buildings full of computers that run websites, apps and cloud services.

Who it affects

The most direct risk is for companies and hosting providers that run data centers using software Canonical maintains. Businesses that rely on those providers could be touched too, even if they never use the software themselves. If you only use a laptop or phone for everyday tasks, there is nothing here you need to change.

What is not known yet

The material available does not name the software, describe how the flaw works, or say whether a fix exists. It also does not say whether anyone has used the flaw to attack a system. Canonical's response has not been described either. Those details will decide how urgent this is.

What to do

Business owners who use cloud or hosted services should ask their provider whether it has dealt with this flaw. The steps in the alert box above cover what to ask and what to watch for. Keep any written answer you get from the provider.