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

Researchers have developed and tested new visual methods to disclose how humans and artificial intelligence collaborate in producing news articles, moving beyond the simplistic labels currently used by publishers [1]. The study, posted to arXiv on January 16, 2026, argues that existing AI disclosure practices in journalism fail to capture the nuance of collaborative workflows [1]. Through 10 co-design sessions, the research team elicited 69 distinct disclosure designs and built four prototypes: a standard textual notice, a role-based timeline, a task-based timeline, and a chatbot interface [1][2]. A subsequent lab study with 32 participants examined how these visualizations and varying collaboration ratios—primarily human versus primarily AI—shaped reader perceptions, gaze patterns, and post-experience responses [1][2]. The findings showed that textual disclosures were the least effective at communicating the nature of human-AI collaboration, while the chatbot visualization provided the most in-depth information [1][2]. The study also revealed that visualizations can skew reader perceptions: role-based timelines amplified the perceived AI contribution in primarily human articles, whereas task-based timelines shifted perceptions toward greater human involvement in primarily AI-generated pieces [1][2]. The authors caution that disclosure design choices can alter public understanding of AI's actual role in news creation [1][2]. The work arrives as newsrooms increasingly integrate generative artificial intelligence into editorial processes. Automated journalism, which encompasses content production, data mining, news dissemination, and content optimization, has expanded in the 2020s with the rise of generative pre-trained transformers capable of producing and summarizing information at scale [3]. While advocates view these tools as a way to free journalists from routine reporting on finance, sports, and elections, critics raise concerns about transparency, accountability, and public trust [3]. The study's focus on visual disclosure methods directly addresses these transparency gaps [1][2]. Broader public discourse has also turned toward the quality of AI-generated media. The term "AI slop"—coined in the 2020s and selected as the 2025 Word of the Year by both Merriam-Webster and the American Dialect Society—describes low-effort, high-volume synthetic content produced for the attention economy [4]. Jonathan Gilmore, a philosophy professor at the City University of New York, described such material as having an "incredibly banal, realistic style" that is easy for viewers to process [4]. The new visualization prototypes offer a potential counterpoint, aiming to give readers a more granular view of how a specific article was assembled rather than leaving them to guess whether it is human or machine-made [1][2].