Archive
Published articles
Every accepted article, with its permanent platform identifier and the editorial reports that produced the decision. Articles are readable in full on this site and are not available for download.
3 of 3 articles · Follow this list by feed
Volume 1, Issue 1 (2026) · Meta-Research on AI in Research · IJAGR/2026/MR/0001
Citation completeness in AI-assisted manuscripts: a corpus study of reference integrity across 4,120 preprints
A. Ngwenya (School of Information Studies, University of Pretoria); H. Lindqvist (Centre for Science Studies, Lund University)
Reference lists are the load-bearing structure of a scientific claim, yet AI-assisted drafting has made plausible-but-nonexistent citations cheap to produce. We assembled a corpus of 4,120 preprints and resolved every listed DOI against the public registry, comparing the resolved record with the cit…
Submitted 12 June 2026 · Accepted 28 August 2026 · Published 10 September 2026 · 2 revision rounds
Volume 1, Issue 1 (2026) · Meta-Research on AI in Research · IJAGR/2026/MR/0002
Agreement and disagreement in automated desk screening: what three independent evaluation remits catch that one does not
R. Osei (Department of Science and Technology Studies, University of Ghana)
Automated screening is usually implemented as a single verdict. We test whether splitting screening into three fixed remits — disciplinary criticism, reference rigour, and contribution — surfaces failures a single generalist pass misses. Using 600 titles, abstracts, and keyword sets with known edito…
Submitted 3 July 2026 · Accepted 2 September 2026 · Published 10 September 2026 · 2 revision rounds
Volume 1, Issue 2 (2026) · AI, Machine Learning & Data Science · IJAGR/2026/AI/0003
Reporting standards for prompt-dependent results: a minimum specification for reproducible AI experiments
M. Haddad (Institute for Applied Computing, American University of Beirut); S. Iwuchukwu (Department of Computer Science, University of Lagos)
Results obtained from instruction-following models are frequently reported without the information needed to reproduce them: exact model version, decoding parameters, prompt text, seeds, and date of access. We specify a minimum reporting set, demonstrate on three published experiments that omitting …
Submitted 21 May 2026 · Accepted 6 September 2026 · Published 15 September 2026 · 3 revision rounds