Source: bleepingcomputer.com
LLMjacking attacks involve threat actors hijacking cloud-hosted large language model (LLM) services to generate fraudulent usage charges. Some victim organizations are losing more than $100,000 per day due to unauthorized AI resource consumption. Google has highlighted these attacks as a significant and growing threat to cloud customers. Attackers typically gain access to cloud credentials and then abuse AI services at scale. The cumulative financial damage from these attacks has surpassed $1 million in cloud AI charges. This type of attack targets the billing infrastructure of cloud AI platforms rather than traditional data exfiltration. Organizations using cloud-based AI services are urged to monitor usage anomalies and secure their credentials.
