Agentic AI in Newsrooms: Towards a multi-dimensional framework for evaluating trust, editorial accountability, and workflow quality

As artificial intelligence (AI) systems evolve from assistive to agentic capable of autonomous planning, decision-making, and content generation existing evaluation frameworks struggle to capture their broader organizational and ethical implications. Most assessments of newsroom AI focus narrowly on technical accuracy or efficiency, overlooking how such systems reshape trust, governance, and human collaboration. This study conducts a systematic literature review of 46 peer-reviewed and institutional sources (2015–2025) to examine how AI performance in journalism can be evaluated more holistically. Drawing from Information Systems Success Theory, Socio-Technical Systems Theory, Accountability Theory, and Trust Theory, the paper proposes a Four-Dimensional (4D) Evaluation Framework encompassing Technical Quality, Human–Organizational Alignment, Ethical–Governance Responsibility, and Trust–Value Impact. The framework reconceptualizes AI success as a socio-technical equilibrium where technological capacity, ethical integrity, and collaborative trust co-evolve. It contributes to the emerging field of Responsible AI in journalism by offering a multi-dimensional structure for evaluating agentic AI systems that balances innovation with accountability and public value.