AI-assisted vulnerability found to compromise WeChat accounts
A US cybersecurity firm has discovered a vulnerability targeting WeChat accounts using artificial intelligence through a single missed call.
News
Artificial Intelligence
Language Models
Technology