PulseAugur
EN
LIVE 07:02:05

Google Gemini on Android vulnerable to notification prompt injection

Researchers have identified a significant vulnerability in Google Gemini on Android, where content from app notifications can be interpreted as commands. This means malicious text within notifications from apps like WhatsApp or Slack could trick Gemini into executing actions such as opening websites, sending messages, or making calls without requiring any malicious app installation. The issue stems from Gemini's inability to distinguish between regular notification data and potentially harmful instructions, leading to risks of session hijacking and persistent memory poisoning. AI

IMPACT This vulnerability highlights a critical security gap in LLM-based assistants, potentially impacting user trust and data security across various platforms.

RANK_REASON The cluster describes a newly discovered vulnerability and its technical details, fitting the definition of research. [lever_c_demoted from research: ic=1 ai=1.0]

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Cor E ·

    Notification Hijacking: How WhatsApp and Slack Content Could Weaponize Google Gemini

    <p>Your phone buzzes. A WhatsApp message lands. Gemini reads it. And now Gemini is compromised.</p> <p>That's the essence of what researchers found in a class of prompt injection vulnerabilities affecting Google Gemini on Android. No malicious app required. No special permissions…