The Hacker News

Google Gemini Prompt Injection Flaw Exposed Private Calendar Data via Malicious Invites

Researchers found an indirect prompt injection flaw in Google Gemini that bypassed Calendar privacy controls and exposed ...

Run one prompt, get multiple AI answers at once—just $80 for lifetime access

Get lifetime access to ChatPlayground AI Unlimited Plan and compare outputs from 25+ top AI models in one powerful interface ...

Reprompt attack hijacked Microsoft Copilot sessions for data theft

Researchers identified an attack method dubbed "Reprompt" that could allow attackers to infiltrate a user's Microsoft Copilot session and issue commands to exfiltrate sensitive data.
eWeek

A ‘Zombie’ AI Attack Shows How ChatGPT Can Leak Your Secrets

Radware’s ZombieAgent technique shows how prompt injection in ChatGPT apps and Memory could enable stealthy data theft ...