Ethical Governance of Artificial Intelligence Across the News Lifecycle

Abstract

Artificial intelligence now supports or shapes news gathering, verification, production, ranking, translation, and audience engagement. These uses can expand reporting capacity, but they also redistribute editorial power and create risks involving privacy, discrimination, false or synthetic evidence, opaque decisions, worker autonomy, and unequal access to reliable news. This conceptual article develops a risk-proportionate framework for newsroom governance by connecting journalism ethics with the principles visible in the Global MOOC on the Ethics of AI and with international guidance and recent research. It argues that ethical performance depends less on a generic label for AI than on the task, data, degree of editorial influence, affected people, and remedies available when harm occurs. The proposed approach requires a public-interest purpose, data and rights review, task-level risk controls, meaningful human editorial authority, context-sensitive disclosure, and continuous monitoring with correction and appeal. Three original figures translate these principles into a news-lifecycle map, a task-risk guide, and an accountability loop. The central conclusion is that newsroom AI should be governed as editorial infrastructure: its use must remain answerable to professional judgment, pluralism, human rights, and the public's ability to question and correct journalism.