add AI security branch

[realmarcin]

LLMs are introducing a lot of new concepts around prompting and prompt architectures. We could even have an entire cybersecurity branch of AIO eg https://greshake.github.io/