NSA tests Anthropic Mythos AI to find software security flaws
Updated
Updated · Bloomberg · Apr 30
NSA tests Anthropic Mythos AI to find software security flaws
9 articles · Updated · Bloomberg · Apr 30
The agency is using the new model to probe vulnerabilities in widely used software, including Microsoft products, according to a US official and another person familiar with the matter.
Officials evaluating Mythos have been impressed by its speed and efficiency in identifying potential weaknesses, the people said.
The testing highlights growing use of advanced AI in cybersecurity, as governments and companies seek faster ways to detect flaws in critical software.
The Pentagon blacklists an AI company, but the NSA uses its tool. What does this reveal about US cyber strategy?
An AI can hack nearly any system. How can we prevent it from becoming the ultimate cyberweapon?