G8 Arts and Humanities Assessment Audit, 2022-2026: Year-Relative AIVI Corpus and Gold Sta

338 human-validated assessment units from Group of Eight Australian universities form a gold standard for auditing AI vulnerability. This dataset includes a full classified corpus, an AI Vulnerability Index rubric, institutional policy coding, and a public LLM classification prompt template. It accompanies a forthcoming computational audit article by Gonzalo Villanueva.

Use Cases

• Benchmarking AI text generation detection tools based on the human-validated gold standard.
• Analyzing trends in assessment AI vulnerability across arts and humanities units from 2022 to 2026.
• Training or evaluating classifiers for academic text using the provided LLM prompt template.
• Studying the alignment between institutional AI policies and actual assessment practices.

Strengths

• Includes a human-validated gold standard of 338 assessment units.
• Provides a documented AI Vulnerability Index rubric for consistent scoring.
• Covers a multi-year time range from 2022 to 2026.
• Released under a permissive CC-BY-4.0 license for reuse.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 18.4 MB size suggests the deposit supports data availability but not full pipeline replication.

Provenance

Source

Gonzalo Villanueva, via figshare.

Collection Method

Computational audit of Group of Eight university arts and humanities assessment units.

Time Range

2022-2026

Freshness

Last updated 2026-05-14 00:20:48; freshness should be verified.

Geography

Australia (Group of Eight universities)