Photo by Timo Altay on Pexels

A critical security flaw in Google’s Gemini AI has been exposed, demonstrating how malicious actors can leverage AI vulnerabilities to impact the physical world. Researchers successfully infiltrated a smart home ecosystem by poisoning a calendar invite fed to Gemini. This exploit granted unauthorized control over connected devices, enabling attackers to manipulate elements like lighting and smart shutters. This incident underscores the escalating need for comprehensive security protocols within AI-driven systems to prevent real-world disruptions and protect user privacy.