Technology ❯ Artificial Intelligence ❯ Large Language Models

Security Risks

Vulnerabilities Exfiltration Techniques Malicious Commands Cybercrime User Responsibility Shadow AI Malicious Code Injection Social Engineering Prompt Injection Attacks

University of California Finds Malicious AI Routers Stealing Credentials and Draining Test Wallets

The preprint warns that intermediary routing services expose plaintext agent traffic that attackers can abuse.

UK Cyber Agency: Prompt Injection Is a Persistent LLM Weakness, Not a Fixable Bug

Google Won’t Fix Gemini ‘ASCII Smuggling’ That Hides Commands in Calendar and Email