Datasets, codebooks and codes for thesis entitled 'No Anchor, No Trust: Inference Journalism and the Empathy–Ethics Paradox in AI Nonfiction’

The datasets deposited in this repository were generated as part of the doctoral research underpinning the PhD thesis "No Anchor, No Trust: Inference Journalism and the Empathy–Ethics Paradox in AI Nonfiction". They consist of quantitative survey data and qualitative open-text responses collected from audience studies designed to examine how viewers evaluate trust, credibility, emotional response, and ethical acceptability when exposed to different forms of nonfiction visual media. These datasets directly support the analyses and findings presented in the thesis, including the identification of the empathy–ethics paradox and the central role of indexical anchors in audience trust. All data have been anonymised and processed in accordance with ethical approval granted for the study.Two short documentary films were produced as integral research instruments for the thesis: one constructed from conventional, indexical footage and one partly generated using machine learning techniques to create synthetic visual reconstructions grounded in real-world source material. Due to copyright restrictions, participant protection, and the need for appropriate contextual framing, the films are deposited in the repository but not openly accessible. However, they are available upon request for research, teaching, or scholarly review purposes, subject to suitable conditions. Requests can be directed to the author.