Generative AI in Organizational Communication: A Mixed-Method Eye-Tracking Study of Content Evaluation, Source Uncertainty, and Human Oversight

Generative AI is increasingly integrated into organizational communication workflows, but organizations still have limited empirical evidence on how recipients evaluate AI-supported communication content when its source is not disclosed. This study examines how AI-generated organizational communication content becomes a decision alternative under hidden-source conditions and what this means for communication management. An exploratory mixed-method eye-tracking experiment was conducted with 20 participants, who evaluated six pairs of organizational communication stimuli, including text-based, image-based, and image-plus-text materials. Each pair contained one AI-generated and one human-created alternative, while the source remained concealed during the initial choice task. Data were collected through paired content-choice tasks, AOI-based eye-tracking with a Tobii Pro Spark 60 Hz screen-based eye tracker, an AI-identification task, and semi-structured post-experiment interviews. Within this six-pair exploratory stimulus set, AI-generated alternatives were selected in 66.7% of participant-by-task decisions, indicating that these specific AI-generated items were not automatically disadvantaged when their origin was hidden. Eye-tracking results showed broadly similar visual attention toward AI-generated and human-created content, while selected alternatives received higher attention than non-selected ones. Interviews revealed that participants evaluated content mainly through perceived quality, structure, professionalism, authenticity, and communicative suitability. The study contributes to research on AI-supported organizational communication by showing that responsible AI use requires human oversight, editorial control, and clear organizational guidelines.