AI Spam Filters Defeated by Decades-Old Text Salting

New AI-powered defenses can be vulnerable to old, low-tech evasion tactics.

Abstract representation of text obfuscation defeating modern filters
AI-generated illustration · Sylvaris

Old trick, new targets

Text salting is a technique where spammers insert invisible or barely visible characters into messages to evade keyword-based filters. The method dates back decades to early email spam wars.

Some LLM-powered email filters are now proving susceptible to these same techniques. The AI systems, trained to understand natural language, can be fooled by character insertions that would trip traditional pattern matching.

Why AI filters miss it

Large language models are designed to be robust to minor text variations, a feature that helps them handle typos and informal writing. That same flexibility becomes a weakness when spammers deliberately introduce noise.

The finding suggests that combining multiple detection approaches — statistical, pattern-based, and AI-powered — may be more resilient than relying on any single method.

sources
more in Security
Upbound breach enabled $13 million in fraudulent Acima leases Stolen customer data was directly weaponized to create fraudulent financial contracts, demonstrating how breach data enables immediate financial crime. Fake job interview delivers malware through Git hooks in take-home coding projects Attackers are weaponizing the technical interview process itself, embedding malicious Git hooks in legitimate-looking coding assignments to compromise developer workstations. South Korea National Diplomatic Academy breach exposes global diplomat data after ten-month intrusion A prolonged breach of diplomatic training infrastructure exposed sensitive personnel data of current and former foreign service officers worldwide, demonstrating the targeting of government educational systems.