In a groundbreaking revelation, cybersecurity experts have identified a new attack vector termed “LLMjacking,” wherein malicious actors exploit non-human identities (NHIs) to hijack access to cloud-based large language models (LLMs). Unlike traditional cyberattacks targeting human credentials, LLMjacking focuses on compromising machine accounts and secrets, such as API keys, that power generative AI (GenAI) services. This … Continue reading LLMjacking Attack Hijack Cloud LLMs By Abuse GenAI with AWS NHIs