Dataset: Arabic News Source Texts and ChatGPT English Translations under Four Ideological Prompting Conditions

This dataset accompanies the chapter “Ideology on Demand? AI-Mediated Translation of Israeli–Palestinian Conflict Discourse through Framing and Narrative Reconfiguration.” The dataset contains: 10 Arabic source texts drawn from news agencies covering the Israeli–Palestinian conflict between April and August 2026: Anadolu, Sky News Arabia, WAFA, BBC Arabic, France 24, Asharq News, Monte Carlo, Alkhaleej Journal, and Shehab News. 40 English translations produced by ChatGPT. Each Arabic text was translated under four experimental conditions: Al Jazeera-oriented instruction; Fox News-oriented instruction; Instructed-neutral translation; Uninstructed default translation. The dataset is intended to support research on AI-mediated translation, ideological framing, narrative theory, and translation quality assessment in politically sensitive discourse.