Cyrillic look-alike characters can silently bypass AI prompt safety filters
A technique called homoglyph evasion allows attackers to bypass AI content filters by substituting standard Latin characters with visually identical Cyrillic or Greek Unicode equivalents. Because…
Read the full articleYou might also wanna read
Invisible Characters as an Attack Vector
Unicode's invisible characters are being weaponized — hiding malicious code in repositories, hijacking AI agents, and bypassing security rev
Attackers use invisible Unicode characters to hide malicious code in GitHub repositories
Unicode that's invisible to the human eye was largely abandoned—until attackers took notice.
arstechnica.com·4mo ago
AEGIS: An Inference-Time Defense Against Visual Synonym Jailbreak Attacks on Text-to-Image Models
Text-to-image diffusion models have achieved high visual fidelity and broad adoption, but remain vulnerable to safety violations when advers
arxiv.org·7d ago
Adversarial Poetry Functions as Universal Jailbreak Technique for Large Language Models
We present evidence that adversarial poetry functions as a universal single-turn jailbreak technique for Large Language Models (LLMs). Acros
arxiv.org·7mo ago
Gibberifier: Tool Uses Invisible Unicode Characters to Block AI Text Scraping
Gibberifier - Block AIs and web scrapers from reading your text with invisible Unicode characters while preserving meaning for humans.
gibberifier.com·7mo ago
AI Jailbreak Technique Exploits LGBT-Related Content Guardrails
🌙 ZetaLib - The only AI Library you need. Contribute to Exocija/ZetaLib development by creating an account on GitHub.
GitHub·2mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.