CoChat

Reimagining writing assessment for the AI era: a systematic review on balancing AI support and authentic skill growth

Open access

The DOI matches its Crossref record; 82 of the 84 references it cites were found in a registry or an index; 121 of the 137 sources its claims cite were checked for support.

Abstract

This systematic review synthesizes empirical evidence from 19 studies examining AI (AI) integration in academic writing assessment across higher education contexts. Following PRISMA 2020 guidelines, the review addresses five interconnected research questions regarding AI tool functionality, stakeholder perceptions, student motivations, institutional implementation gaps, and assessment redesign strategies. Findings reveal that generative AI platforms, particularly ChatGPT (88.8% adoption) and Grammarly (67.4%), dominate across writing stages, with students strategically deploying tools based on cognitive complexity levels. Stakeholder perspectives diverge significantly: students value productivity enhancement and accessibility, faculty prioritize academic integrity concerns, and administrators adopt cautious openness with disproportionate resource allocation favoring faculty guidance (67%) over direct student support (17.8%). Student motivations for inappropriate use stem from structural vulnerabilities including time pressures, ethical ambiguity, and institutional policy voids rather than malicious intent. Geographic inequities persist, with Global South institutions facing infrastructure limitations that amplify educational disparities. Promising pedagogical and assessment innovations include process-oriented assessment, oral defenses, multimodal composing, and explicit AI literacy curricula—each targeting the redesign of writing evaluation practices to balance AI assistance with authentic skill development. The review concludes that effective integration requires moving beyond detection-focused approaches toward comprehensive pedagogical redesign grounded in equity, authentic learning, and development of uniquely human capacities for critical judgment and ethical reasoning.

Overview

CoChat Summary

TL;DR

This PRISMA 2020 systematic review synthesizes 19 empirical studies on AI integration in higher-education academic writing assessment. It finds stratified AI tool use, divergent stakeholder views, structural drivers of inappropriate use, geographic inequities, and assessment redesign strategies intended to balance AI support with authentic skill growth.

Key findings

  • ChatGPT dominated reported use at 88.8%, followed by Grammarly at 67.4%; Canva (57.3%) and Quizlet (30.3%) were reported in specific contexts. Tools were grouped into generative AI, automated writing evaluation, translation, and multimodal/interactive categories, with LLMs used more for higher-order planning/drafting and AWE tools for revision/editing.
  • Stakeholder perspectives diverged: students valued productivity and accessibility, faculty prioritized integrity concerns, and administrators adopted “open but cautious” approaches. A resource gap was reported: 67% of guidance targeted faculty education while only 17.8% provided direct student frameworks.
  • Inappropriate AI use was attributed less to malicious intent than to structural vulnerabilities including time pressures, ethical ambiguity, and institutional policy voids; “metacognitive laziness” emerged when assessment emphasized final products over learning processes.
  • Geographic inequities persisted, with Global South institutions facing infrastructure limitations and research concentrated in Global North contexts. Promising redesign strategies included oral defenses, process-oriented assessment, multimodal composing, project-based learning, scaffolded assignments, and explicit AI literacy curricula.

Methods

  • PRISMA 2020 systematic review searching Web of Science, Scopus, ERIC, PsycINFO, and Google Scholar for peer-reviewed English literature published January 2020–December 2025, using three thematic search clusters and forward/backward citation tracking.
  • Inclusion required empirical research on AI in higher-education academic writing assessment, including student, faculty, or institutional perspectives; exclusions covered K-12-only studies, non-writing AI applications, purely theoretical pieces, dissertations/theses/preprints, non-English publications, and pre-2020 automated essay scoring without generative AI context.
  • Screening began with 1,756 database records plus 18 citation records; after removing 385 duplicates, 1,389 records were screened, 1,251 were excluded, 138 full texts were assessed, 119 were excluded, and 19 studies were included. Inter-rater reliability was Cohen’s κ = 0.87; heterogeneity led to narrative synthesis, with quality appraisal using adapted Newcastle–Ottawa Scale, CASP, and MMAT.

Limitations

  • The paper states that the 19-study corpus is small; evidence relies heavily on convenience sampling and self-reported data; longitudinal designs are absent; geographic concentration limits transferability; rapid AI evolution creates temporal instability; and publication bias favors positive-utility findings, with possible time-lag bias and marginalization of student voices.
  • Evident from the text: the review uses narrative synthesis rather than meta-analysis because of heterogeneous designs, contexts, populations, and outcomes, and quality ratings were applied descriptively rather than to exclude studies, so conclusions rest on a methodologically varied evidence base.

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

Integrity signals

Registry

Crossref · 10.3389/fpsyg.2026.1809174

Cites

82 of 84 references verified2 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 PRISMA 2020 flow diagram. Systematic review flow diagram adapted from (Page et al., 2021).
Figure 1 PRISMA 2020 flow diagram. Systematic review flow diagram adapted from (Page et al., 2021).CC BY · PubMed Central

Verification

Claims the paper makes

68 Valid · 9 Not valid · 5 Not checked · 23 Own results and interpretations

54 more claims · 42 with their cited sources checked

Create a free account to see all 105 claims and to verify the 2 references still open.

See all claims — free account
  • It may state two separate claims as one. 34
  • Probably not a substantial claim. 17

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

Show all 84 references

Research works printed but not linked: 2

  1. Anderson, 2001, A Taxonomy for Learning, Teaching, and Assessing: A Revision of Bloom's Educational Objectives
  2. Talandis, 2025, Moving JALT into the Future: Opportunity, Diversity, and Excellence

Metadata

Authors
Abdellatif, M. S., Alshehri, M. A., Lamouchi, A., Rahmath, M., & Nemt-allah, M. A.
Affiliations
  • Prince Sattam Bin Abdulaziz University (SA)
  • King Saud University (SA)
  • Al-Azhar University (EG)
Corresponding author
Nemt-allah, Mohamed Ali
Venue
Frontiers in Psychology
Published
2026-05-26
Volume
17
Pages
1809174-1809174
ISSN
1664-1078 (ISSN-L)
OpenAlex
W7162420461
Cited by
0
References
80
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Psychology (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Academic integrity and plagiarism
  • Social Sciences › Education › Student Assessment and Feedback
Keywords
academic writing, generative AI tools, ChatGPT, Grammarly, AI literacy, academic integrity, student motivation, stakeholder perspectives, assessment redesign, process-oriented assessment, multimodal composition, AI detection, educational equity, higher 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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Reimagining writing assessment for the AI era: a systematic review on balancing AI support and authentic skill growth. https://app.cochat.ai/papers/10.3389/fpsyg.2026.1809174
The paper
Abdellatif, M. S., Alshehri, M. A., Lamouchi, A., Rahmath, M., & Nemt-allah, M. A. (2026). Reimagining writing assessment for the AI era: a systematic review on balancing AI support and authentic skill growth. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1809174

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.