AI-generated readings lacked internal traceability: only about 0.80% of pages contained APA-style citations, URLs appeared on about 0.53%, and DOI strings were essentially absent¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? Across the 4,487-page corpus, only about 0.80% of pages contained APA-style in-text citations, about 0.53% contained URLs, and DOI strings were essentially absent, leaving most claims difficult to audit from within the artifact. → Sidorkin 2026 q2 i? About 1.03% of pages contained over-specific institutional claims pairing Sacramento State or Sac State with assertive policy verbs, without verifiable citation. → Sidorkin 2026 (2)

Evidence¶

Sidorkin 2026¶

Sidorkin, A. M. (2026). From One-Size Texts to Tailored Readings: Student Experiences with AI-Generated Course Materials. Open Praxis, 18(3), 506–522. https://doi.org/10.55982/openpraxis.18.3.1141

q2 · i? · design · r2

Automated page-level screen over the full 4,487-page corpus for citation signals. The article reports "only about 0.80% of pages contain APA-style in-text citations" and that most claims remain difficult to audit from within the artifact itself.

"Across the 4,487 pages, only about 0.80% of pages contain APA-style in-text citations in the form “(Author, year)” or similar. URLs appear on about 0.53% of pages, and DOI strings are essentially absent."

Sidorkin 2026 (2)¶

Sidorkin, A. M. (2026). From One-Size Texts to Tailored Readings: Student Experiences with AI-Generated Course Materials. Open Praxis, 18(3), 506–522. https://doi.org/10.55982/openpraxis.18.3.1141

q2 · i? · design · r2

The same automated screen identified institution-specific assertive statements. The article calls this "the most consequential failure mode in an instructional setting, because specificity creates a higher cost of being wrong than generic content."

"About 1.03% of pages contain patterns pairing “Sacramento State” or “Sac State” with assertive policy verbs."

Discussion¶

Related Claims¶

• Interest tailoring was structurally present in AI-generated readings: logs averaged 52.24 tailoring markers per 10,000 words and showed moderate divergence (cosine similarity 0.50–0.61)
• Standardized repair and source-verification gating produced structurally complete, source-traceable feedback for most generated messages — related
• Students showed bounded trust and uneven verification behavior: 23 of 837 student turns (about 2.7%) contained explicit risk-aware moves — related