Anthropic's Mythos AI helped find macOS flaws reported to Apple
Technology
Anthropic's Mythos AI helped researchers spot some serious security holes in Apple's macOS, raising fresh worries about how smart AI tools could be used for cyberattacks.
The team at Palo Alto used Mythos to uncover the flaws and reported everything straight to Apple, according to The Wall Street Journal (May 15, 2026).
Mythos chained exploit needed human guidance
Mythos pulled off a "chained attack" by linking two minor bugs together, letting it mess with system memory and sneak into restricted parts of the device.
Still, human experts had to guide the process and test out the attack.
Apple says it is reviewing the findings and takes vulnerability reports seriously.
This news comes as concerns grow about AI-powered hacks, especially after a recent incident at Google.