AI Spots Major Security Flaws in Internet Encryption
An AI model named Mythos recently scanned complex security codes. It found a way to break a new type of encryption that experts had spent two years reviewing. The computer program discovered these flaws in just 60 hours.
This experiment cost about 100,000 dollars to run. While the specific encryption tested is not yet widely used on the internet, the results are a big deal. They prove that AI can now find deep mathematical errors much faster than humans can.
Anthropic says this is a warning for the future of online safety. We need to prepare for a time when AI might be used to attack our core data systems. Security experts will now have to find new ways to protect our information.
Comments (0)
No comments yet. Be the first!
More AI news
NewsAmazon Shifts Focus to New AI Research
Amazon is stopping active work on its current Nova AI models to focus on a new project.
NewsNvidia Backs Ilya Sutskever’s New AI Lab
Nvidia is investing heavily in Safe Superintelligence to support its upcoming AI projects.
NewsNvidia worker detained in China chip smuggling probe
Taiwanese officials have arrested an Nvidia employee over the illegal shipment of high-end AI servers to China.