Prompt Injection
-
techInstagram AI Chatbot Exploit Exposes Critical Security Flaw in Meta's Automation
Hackers used prompt injection to trick Meta's AI chatbot into handing over high-profile Instagram accounts, raising concerns about AI safety and the pace of automation at the social media giant.
-
techAI Prompt Injection Attacks: The Security Flaw That Won't Be Fixed
Prompt injection exploits a core flaw in LLMs: they can't distinguish instructions from data. Experts say it may never be fully patched.
-
techCertiK CEO Warns Mass AI Agent Deployment is a Security 'Disaster'
Ronghui Gu cautions that unisolated AI agents pose critical security risks, from data theft to automated financial scams.
-
techGoogle Patches Critical AI Coding Tool Vulnerability: Prompt Injection Allowed Code Execution
Google's Antigravity AI coding platform had a flaw allowing attackers to execute malicious code via prompt injection, bypassing security measures. The issue is now fixed.
-
techGoogle DeepMind Maps Six New AI Agent Traps Turning Web Into Adversarial Hunting Ground
Google researchers identify six categories of engineered traps designed to hijack autonomous AI agents, highlighting a critical accountability gap in liability for resulting crimes.
-
techOpenAI Acquires Promptfoo to Bolster AI Application Testing and Security
OpenAI acquires Promptfoo, a startup specializing in AI application testing and security evaluation, to enhance its Frontier platform.