A structured process that security professionals can use to find Indirect Prompt Injection vulnerabilities in LLMs and produce POCs
This is amazing, we need to test this and do it constantly.
I want to test this for IoTs and robots because they also work and, in the vast majority of cases, will function with LLM.
Great point! If AI gets embedded into IoT devices some very interesting attack paths could occur
This is amazing, we need to test this and do it constantly.
I want to test this for IoTs and robots because they also work and, in the vast majority of cases, will function with LLM.
Great point! If AI gets embedded into IoT devices some very interesting attack paths could occur