# AI Spam Filters Fooled by Old-School Text Salting Techniques Cyber attackers bypass email filters using hidden text manipulation methods, reveals Barracuda research. - **Source:** [www.theregister.com - Articles (theregister.co.uk)](https://brifing.si/source/e6d4ec13-4d0e-407a-b96d-39609d76c3b1) - **Published:** 2026-07-17 - **Topic:** [Software & Engineering](https://brifing.si/c/engineering) - **Original article:** - **On Brifing:** https://brifing.si/s/49014ce7-440e-49d3-87fb-15eddfd0d3fe ## What is Text Salting? Attackers add random benign words to malicious emails to deceive AI filters, making scanning systems perceive messages as normal. Barracuda detected over 1 million phishing attacks using this method since April 2023. ## Digital Adaptation of Traditional Tactics Text salting isn't new—it's bypassed traditional email security for years. Now machine learning and LLM-based tools are also vulnerable. Attackers primarily use three variation techniques to hide their tracks from automated scanners. ## How Concealment Techniques Work CSS cropping shrinks the visible window to hide filler text. Text manipulation moves suspicious content off-screen. Zero-font techniques insert machine-readable but human-invisible misleading words between phishing content. ## Why AI Systems Are Vulnerable LLMs typically process email text and source code plainly without understanding visibility context. Barracuda notes these systems can be trained for detection, but most tools don't incorporate this capability by default. ## Enterprise Protection Strategies Barracuda recommends layered email security: checking sender reputation, authentication results, embedded URLs, and HTML rendering differences instead of relying solely on keyword detection. Removing AI spam filters entirely may be considered. ## The Need for Continuous Adaptation These attacks demonstrate that technological advancements alone aren't sufficient. Organizations must continuously update security strategies with multi-layered defenses. Current AI solutions lack human-level contextual understanding for reliable detection. ## Sources covering this story - https://www.theregister.com/security/2026/07/17/ai-spam-filters-are-getting-suckered-by-old-school-text-salting/5274434 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._