AI-Developed Worm Exploits WeChat Accounts via Voice Calls

US security firm Calif identified a critical vulnerability in Tencent's WeChat that allowed an AI-developed exploit, 'WeWorm', to hijack accounts via unanswered voice calls.

Tencent deployed a server-side patch in late August to fix the flaw, confirming that there was no evidence of the vulnerability being exploited in the wild.

The incident has prompted calls for increased bilateral cooperation between the US and China on AI safety, as researchers noted the exploit required minimal human engineering.