
Microsoft has identified a sophisticated phishing campaign using 'ASCII smuggling' to bypass AI-driven email security filters and deliver malicious content.
The technique involves using invisible Unicode characters to mask phishing links, allowing attackers to evade standard detection protocols used by major email providers.
Security researchers are now updating prompt injection detectors to identify these patterns, as the threat has evolved from AI manipulation to widespread phishing.