CoChat

Human-in-the-loop assessment with AI: implications for teacher education in Ibero-American universities

Open access

The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index; 5 of the 12 sources its claims cite were checked for support.

Abstract

This scoping review examines how artificial intelligence (AI) reshapes assessment in Ibero-American higher education and specifies the teacher-training capacities and ethical safeguards needed for responsible adoption. Guided by PRISMA procedures and an eligibility scheme based on PPCDO (Population–Phenomenon–Context–Design–Outcomes), we searched Scopus and screened records (2015–2025; English/Spanish/Portuguese), yielding 76 peer-reviewed studies.

Overview

CoChat Summary

TL;DR

This scoping review of 76 peer-reviewed studies (2015–2025) examines how AI reshapes assessment in Ibero-American higher education and what teacher-training capacities and ethical safeguards are needed. It proposes a function-by-purpose taxonomy of assessment technologies and five competency clusters, concluding that systems are technically ready but pedagogically under-specified.

Key findings

  • From 11,938 initial Scopus records, filtering and sampling reduced the corpus to 76 studies; AI-labeled tools accounted for 8.9% (n = 23) overall, while LMS-based assessment appeared in 17.4% (n = 45) of references, with Moodle in 35.6% (n = 16) of LMS mentions.
  • Interactive quizzes/tests represented 44.8% (n = 13) of assessment tools in that subsection, with ≈70% including interactive elements; rubrics/portfolios were 20.7% (n = 6); video resources 26.9% (n = 18); VR/AR 0.8% (n = 2); programming tools 2.7% (n = 7).
  • Exploratory correlations linked evaluation platforms with continuous feedback (r = 0.83, p = 0.002), simulators with practical skills (r = 0.85, p = 0.001), augmented reality with immersion (r = 0.81, p = 0.001), virtual platforms with time management (r = 0.78, p = 0.002), and authoring tools with personalization (r = 0.74, p = 0.003).
  • The review specifies five teacher-education competency clusters: feedback literacy with AI; rubric/item validation and traceability; data interpretation and fairness; integrity and transparency in AI-involved assessment; and orchestration of platforms plus moderation/double-marking when AI assists scoring.

Methods

  • Scoping review guided by PRISMA procedures and a PPCDO eligibility scheme (Population–Phenomenon–Context–Design–Outcomes), using Scopus as the sole primary database and screening peer-reviewed articles/reviews from 2015–2025 in English, Spanish, or Portuguese.
  • Screening proceeded from 11,938 records to 4,617 after subject/document filters, then a probabilistic sample of 355, then 106 after title/abstract screening, yielding 76 studies for extraction.
  • Synthesis combined qualitative thematic analysis, quantitative descriptors/frequencies, and exploratory Pearson correlations between tool families and reported advantages, with significance set at p < 0.05; AI was coded only when primary studies explicitly reported generative, predictive/recommender, or automated scoring/feedback features.

Limitations

  • Stated in the paper: only Scopus was reviewed, so complementary studies in other databases may be missed; because technology research advances rapidly, readers should consider the reviewed timeframe and compare with more recent work.
  • Evident from the text: the correlational analysis is bivariate and does not establish causality, and several outcomes (e.g., motivation) are broader than assessment per se.
  • Evident from the text: the single-database design and sampling strategy may constrain coverage and generalizability, although the authors report forward-backward citation chasing to address single-database bias.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/feduc.2025.1710992

Cites

26 of 26 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

5 Valid · 6 Not valid · 6 Not checked · 82 Own results and interpretations

33 more claims

Create a free account to see all 99 claims and to verify the 1 reference still open.

See all claims — free account
  • It may state two separate claims as one. 45
  • Probably not a substantial claim. 28

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Research works printed but not linked: 1

  1. Chaparro-Martínez, 2016, Gestión de la información: Uso de las bases de datos scopus y web of science con fines académicos., Univers. Ciencia Tecnol., 20, 166

Cited by

See all 29

Metadata

Authors
Fajardo-Ramos, D. C., Chiappe, A., & Mella-Norambuena, J.
Affiliations
  • Universidad de La Sabana (CO)
  • University of the Americas (CL)
Corresponding author
Chiappe, Andrés
Venue
Frontiers in Education
Published
2025-10-23
Volume
10
ISSN
2504-284X (ISSN-L)
OpenAlex
W4415457454
Cited by
29
References
24
Citations by year
2026: 29
FWCI
30.15
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Education (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Universidad de La Sabana
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Student Assessment and Feedback
  • Social Sciences › Education › Educational Technology and Pedagogy
Keywords
AI-assisted assessment, teacher professional development, Ibero-American higher education, AI-assisted grading, generative AI, adaptive testing, algorithmic feedback, explainability, data stewardship, audit trail, rubric validation, duplicate marking, AI ethics in education

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.