Invisible Unicode tricks fuel spam as ASCII smuggling widens its reach
An Ars Technica piece explains how “ASCII smuggling” uses invisible Unicode tag characters to hide malicious prompts in emails, enabling spammers to evade filters and trick AI detectors. Microsoft Defender logged a sharp spike in ASCII-smuggling signatures in early 2026, highlighting that traditional keyword searches and even some ML-based spam detectors can miss these hidden tokens unless filters are updated to account for invisible characters or use OCR-like analysis.
