Prioritizing Generative Artificial Intelligence Co-Writing Tools in Newsrooms: A Hybrid MCDM Framework for Transparency, Stability, and Editorial Integrity

Abstract

The rapid integration of generative artificial intelligence (AI) into newsroom workflows has transformed journalistic writing. Still, selecting reliable co-writing tools remains a multi-criteria challenge as it involves technical, ethical, and economic trade-offs. This study develops a hybrid multi-criteria decision-making (MCDM) framework that integrates the Measurement of Alternatives and Ranking according to the Compromise Solution (MARCOS) model with Entropy, CRITIC, MEREC, CILOS, and Standard Deviation objective weighting methods fused through the Bonferroni operator to reduce subjectivity and enhance robustness. Nine generative AI tools, including ChatGPT, Claude, Gemini, and Copilot, were evaluated against sixteen benefit- and cost-type criteria encompassing accuracy, usability, transparency, risk, and scalability. The decision matrix was normalized and benchmarked against ideal and anti-ideal profiles. The MCDM model was validated through correlation and sensitivity analyses using Spearman’s and Kendall’s coefficients. The results indicate that Gemini and Claude achieved the highest overall performance due to superior factual accuracy, transparency, and workflow integration, while ChatGPT demonstrated high linguistic versatility. The hybrid model achieved a stability index above 0.9 across perturbation scenarios, confirming its consistency and reliability. Overall, the proposed MARCOS–objective weight framework provides a mathematically transparent and reproducible decision protocol for newsroom technology evaluation, supporting evidence-based selection of generative AI co-writing systems.

Ähnliche Arbeiten