LLMjacking attacks drive $1M+ cloud AI fraud charges per victim

Google reports a surge in LLMjacking attacks, where threat actors steal cloud credentials to run large language models, generating fraudulent charges exceeding $100,000 per day for some organizations. The attacks exploit misconfigured cloud environments and are increasingly automated, targeting AI workloads. This marks a notable escalation in cloud AI abuse, with financial and operational risks for enterprises.

AI disclosure: Model-written summary of the source, not a verified fact. Check the sources. How Mosaic works