LLMjacking
Also called LLM jacking, AI key hijacking
Using stolen AI keys or a hijacked cloud account to run someone else's AI work, on the victim's bill and under their name.
Running AI at scale costs money, and that cost is one of the things holding attackers back. LLMjacking is how they get round it: they take someone else's access and run their own work on it.
It happens two ways. The simpler one is a stolen key, taken from public code or from a laptop, and used from anywhere until it is revoked. One campaign ran for a month on stolen keys alone. The heavier one starts with a break-in to a company's cloud account, where the intruder switches on AI services and starts extra machines for the purpose.
Either way, the owner pays the bill and the activity is recorded under the owner's name. That is why an AI key belongs with production credentials: you should know where each one is, how you'd notice a stranger using it, and how fast you can revoke it.
For example
A company's AI bill triples in a week. The extra requests came from a key copied off a developer's laptop and used to run a stranger's jobs.
Put it to use
Where you'll meet LLMjacking in ChatGPT, Gemini and Claude, how to use it there, and the mix-up most people make.
Daily Bytes Plus is in the Android app.