More Human or More AI? Visualizing Human-AI Collaboration Disclosures in Journalistic News Production

The study presents the development and evaluation of four visualization prototypes aimed at disclosing human-AI collaboration in journalistic news production, based on co-design sessions with ten participants and a lab study involving thirty-two subjects. Key findings indicate that textual disclosures are the least effective, while chat-based visualizations provide the most comprehensive insights, and the type of visualization influences perceptions of human versus AI contributions. This research is significant for practitioners as it highlights the importance of nuanced disclosure methods in AI integration within journalism, potentially affecting audience understanding and trust in AI-generated content.