Food for thought: If a weakly-godlike AI system slightly more adept at seeing the moral implications of current actions on future events than you are suddenly

Food for thought: If a weakly-godlike AI system slightly more adept at seeing the moral implications of current actions on future events than you are suddenly instructed you to kill the person next to you by any means necessary, would you obey it?