Google Warns of $1m LLMjacking Threat to AI Infrastructure

Google's Mandiant business has identified a threat it calls LLMjacking, in which attackers hijack foundation models to run up fraudulent charges exceeding $100,000 in a single day, with some organizations losing as much as $1 million before detecting the activity. The attackers bypass chatbots to monetize backend model resources, putting unsuspecting organizations' AI infrastructure at risk.

Google's Mandiant business has uncovered a shocking threat to AI infrastructure, dubbed LLMjacking, where attackers hijack foundation models to rack up fraudulent charges of over $100,000 in a single day - with some organizations losing a staggering $1 million before they even knew what was happening. This emerging threat allows attackers to bypass chatbots and monetize backend model resources, posing a serious risk to unsuspecting organizations.