1M+ Emails Use Hidden Text to Dupe AI Security Filters

More than one million emails have been detected using hidden text or 'text salting' to bypass AI-powered security filters. The technique exploits how LLMs process visual elements differently than traditional filters.