Ethical‑AI Literacy Pilot—Bias‑Detection & Reflection Data (Version 1.0)

Dataset and open‑license teaching materials for “Ethical‑AI Literacy Through Scenario‑Based Tasks: A Six‑Week B2 Pilot” (AsiaTEFL 2025 proceedings, Sections 4‑5). CSV summaries• bias_test_summary_verified_v1.csv – pre/post means, SE, 95 % CI for the 12‑item bias‑detection test (N = 50).• reflection_depth_summary_v2.csv – baseline vs post‑test medians plus Wilcoxon W = 0, r = 0.88, p = 0.00031 (39 matched logs).• qualitative_themes_summary_v1.csv – counts and percentages for four reflection‑log themes.• teacher_workload_survey_v1.csv – item‑level means ± SD for the 5‑teacher workload survey. Toolkit files (CC BY‑NC 4.0)• scenario_bank.xlsx – 40 scenario‑based tasks (blank, reusable).• bias_test_forms.docx – Forms A & B with answer keys (student‑ID line removed).• cue_card.pdf – pocket Source–Bias–Impact cue card (no institutional branding).• reflection_log_form.csv – six‑item blank reflection log template.• reflection_rubric.pdf – 4‑point analytic rubric. All files are fully anonymised; no student‑level or teacher‑identifiable data are included. The dataset supports reproducibility while complying with the manuscript’s data‑availability statement and institutional ethics approval (permit code withheld for blind review).