Google flags criminals hijacking corporate AI accounts for fraudulent charges
Google's threat intelligence team has documented criminal services that resell stolen access to businesses' AI models.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
Google's threat intelligence team uncovered underground services that resell compromised access to enterprise AI models. Attackers use stolen credentials to run large language models on company accounts, leaving victims with large cloud bills. Google says the charges reach $1 million at some organizations.
Why it matters
Victims are left paying for AI usage they never ordered, with costs that can run into the hundreds of thousands or more. The attacks are reported to be surging.
Timeline
- Nov 3, 6:52 AM ET: Google's threat intelligence team uncovered underground services reselling compromised access to enterprise AI models, leaving victims with fake cloud bills.
- Nov 3, 10:41 AM ET: Google detailed LLMjacking attacks that hijack enterprise AI resources, with fraudulent cloud charges reaching $1 million at some organizations.
- Nov 4, 12:42 AM ET: Reports said LLMjacking is surging as attackers hijack AI access and push costs onto companies.
Coverage
- LLMjacking Attacks Steal Corporate AI Access, Run Up $100,000 Daily Bills (Hyun-Seon)
- LLM jacking surges as attackers hijack AI access and dump costs on companies (Lee Jae-eun)
More from Technology & AI
- China puts humanoid robots to work serving food and sorting parcels
- Newsom proposes sweeping AI safety rules for Washington
- Google releases Gemini 4 Neon AI model with million-token context
- Microsoft and Intel widen their cloud chip partnership
- Google DeepMind rolls out Genie 4 to select subscribers
- AI helped AMD shorten chip design work, OpenAI says