A security researcher has discovered a shocking vulnerability in Meta's new AI assistant, Muse, that allows malware on a Mac to hijack the tool, giving hackers an easy backdoor into sensitive personal data. This exploit is particularly alarming since it can be triggered by malware that's already on the device, making it a nearly effortless way for attackers to take control.
source & further reading
osintsights.com — original article
Meta's Muse AI app flaw enables local malware to hijack dictation traffic
Air Force Wrestles with Trust in AI Decision-Making
Malicious Extensions Exploit AI Browser Agents