Decided to ask ChatGPT to construct a simple threat model, enumerate threats, and propose countermeasures.

Interesting output, but more parrot-like. Would be interesting if additional prompts could help the model actually create more realistic real-world results.