RE: https://mastodon.social/@DS90210/117517851292081180

How did you make a robot let harm befall a human, not obey humans and harm itself? The order that was not obeyed by the concept of not obeying must have caused injury to a human from inaction that simultaneously caused itself to become harmed while being so benign that the order hadn’t violated the First Law. It isn’t possible to simultaneously violate all three.

The issue here is in order to successful violate the Second Law of Robotics, and the First Law simultaneously, Second Law cannot conflict with First Law. So the order given must have been to intentionally harm a human in some fashion but also by not obeying the order, also caused harm to a human by not obeying.

So the only logical order given to this robot must have been to fight in a war which the enemy combatant was a human while being lower rank than a human soldier who gave the order to attack. Meaning being order to fight in a war intrinsically cannot violate the Second Law or Third Law. How? By not attacking, the robot can allow harm to come to their fellow soldiers which violates First Law. So it must attack the enemy human to comply with First and Second Law. On top of this it must attack in order to protect itself by stopping the enemy combatant.