Do we trust artificial intelligence agents to mediate conflict? Not entirely: New study says we'll listen to virtual agents except when goings get tough