No validated individual-level instrument in the focal corpus tests the full combination of agent-control behaviors for tool-using AI agents¶

Claim · All claims
Evidence · 1 study (2 entries) · quant-synthesis r1 · 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? No validated individual-level instrument in the focal corpus tests the full combination of scope changes, permission decisions, recovery from faulty state, state isolation, independent review, and evidence-based closure. → Daniele Verí 2026 q2 i? Nearby sources cover parts of the target: human-agent teaming metrics measure team behavior rather than individual proficiency, and a task-oriented Navy scenario assessment is a context-specific preprint. → Daniele Verí 2026 (2)

Evidence¶

Daniele Verí 2026¶

Daniele Verí. (2026). Beyond AI Literacy: A Structured Review and Exploratory Meta-Analysis of Measures for Competent Generative-AI Use. arXiv. https://arxiv.org/abs/2609.15624 doi:10.48550/arxiv.2609.15624

q2 · i? · quant-synthesis · r1

Corpus-bounded coverage finding from the review's conclusion, consistent with its operational-agent gap section listing untested behaviors such as selecting the smallest permission and action surface and separating maker and reviewer roles.

"No validated individual-level instrument in the focal corpus tests the full combination of scope changes, permission decisions, recovery from faulty state, state isolation, independent review, and evidence-based closure considered here."

Daniele Verí 2026 (2)¶

Daniele Verí. (2026). Beyond AI Literacy: A Structured Review and Exploratory Meta-Analysis of Measures for Competent Generative-AI Use. arXiv. https://arxiv.org/abs/2609.15624

q2 · i? · quant-synthesis · r1

Review's assessment of adjacent sources: Dorneich et al. (2023) operationalize team behaviors in gameplay, while Bogart et al. (2025) show a scenario task was more informative for applied AI literacy than generic tests, but that work is a preprint without a general normed scale.

"Their unit of analysis is team behavior in gameplay, not individual workplace proficiency. Bogart et al. move closer to occupational assessment: in a US Navy robotics training context, a scenario task designed to simulate job use was more informative for applied AI literacy than generic tests adopted or developed for the program [Bogart et al., 2025]."

Discussion¶

Related Claims¶

• Agentic AI invalidates the three assumptions (evaluation, reversibility, control) on which existing AI literacy frameworks rest — related
• Current AIED rests on five faulty assumptions that risk learner dependence, isolation, and reduced flourishing — related