According to the course text, “if you don’t try to help the people who will use your system to use it securely, they will probably fail”.
In your own words, why do you think modeling for systems that involve people is harder than other threat modeling?