Google Gemini AI Tricked Into Leaking Calendar Data via Meeting Invites

Researchers at Miggo Security demonstrated a semantic prompt-injection attack against Google Gemini using ordinary calendar invites: malicious instructions hidden in an invite description cause Gemini to read private meetings and create a new calendar event containing sensitive summaries, effectively exfiltrating data without user interaction; Google patched the specific flaw but the report warns that AI assistants’ helpful behavior creates a persistent attack surface for language-based exploits.