Which of the following is a primary concern when developing AI systems to ensure ethical outcomes?