Over 1M phishing emails use hidden text to evade AI filters

Over 1 million phishing emails have used hidden text (text salting) to bypass AI security filters. The technique renders LLM-based detection ineffective, allowing malicious emails to reach inboxes.