The DOI matches its Crossref record; 29 of the 29 references it cites were found in a registry or an index; 7 of the 24 sources its claims cite were checked for support.
Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes
Open accessAbstract
Introduction: Writing with generative artificial intelligence (GenAI) can improve a supported product without necessarily strengthening the competence that remains after support is removed. This study examined that distinction through a layer-sensitive account of cognitive offloading. Methods: In this eight-week quasi-experimental classroom study, 180 Chinese undergraduates from six intact English academic-writing classes entered the study and 168 completed the protocol. Two classes were assigned to each of three conditions: no-AI writing, bounded AI support with compulsory reflection, and open AI collaboration. Students completed a baseline task, three intervention assignments, and a supervised Week 8 independent no-AI near-transfer task. Results: Open AI collaboration had the highest observed supported-writing mean ( M = 4.02), although the class-clustered open–bounded contrast was imprecise (adjusted difference = 0.21, 95% CI [0.06, 0.36], wild-cluster p = 0.064). At Week 8, the bounded-support condition showed higher adjusted outcomes than open collaboration in writing quality (difference = 0.27), higher-order thinking (0.35), argument depth (0.42), and independent revision quality (0.39); exact wild-cluster p values ranged from 0.050 to 0.063. Within the 112 AI-exposed students, open collaboration involved more prompts, greater text incorporation, less self-generated text, and deeper idea and reasoning offloading. An adjusted associative decomposition showed that open collaboration predicted higher aggregate offloading ( a = 0.77, SE = 0.06), which was associated with lower Week 8 higher-order thinking ( b = −0.45, SE = 0.08); the bootstrap indirect estimate was −0.34, 95% CI [−0.48, −0.21]. Self-regulated writing attenuated, but did not eliminate, this negative association. Discussion: Because conditions were allocated by only six intact classes and the bounded condition combined delegation restrictions with compulsory reflection, the results are interpreted as classroom-level, mechanism-consistent associations rather than definitive causal effects. The findings distinguish performance achieved with GenAI from competence demonstrated independently after support is removed.
Overview
CoChat Summary
TL;DR
In an eight-week quasi-experimental classroom study, open GenAI collaboration produced the highest supported writing quality, but bounded AI support with compulsory reflection showed better independent no-AI Week 8 outcomes. The results are framed as classroom-level, mechanism-consistent associations linking deeper cognitive offloading and reduced self-generated practice to weaker independent higher-order thinking.
Key findings
- Open AI collaboration had the highest observed supported-writing mean (M = 4.02), but the class-clustered open–bounded contrast was imprecise (adjusted difference = 0.21, 95% CI [0.06, 0.36], wild-cluster p = 0.064).
- At Week 8, bounded support exceeded open collaboration in adjusted writing quality (difference = 0.27), higher-order thinking (0.35), argument depth (0.42), and independent revision quality (0.39); exact wild-cluster p values ranged from 0.050 to 0.063.
- Among 112 AI-exposed students, open collaboration involved more prompts per task (13.92 vs. 7.58), greater AI text incorporation (42.7% vs. 19.4%), less self-generated text (0.56 vs. 0.78), and deeper idea and reasoning offloading.
- An adjusted associative decomposition showed open collaboration predicted higher aggregate offloading (a = 0.77, SE = 0.06), which was associated with lower Week 8 higher-order thinking (b = −0.45, SE = 0.08); the bootstrap indirect estimate was −0.34, 95% CI [−0.48, −0.21]. Self-regulated writing attenuated but did not eliminate this negative association.
Methods
- Eight-week quasi-experimental classroom study with 180 Chinese undergraduates from six intact English academic-writing classes; 168 completed the protocol, with two classes assigned to each of three conditions: no-AI writing, bounded AI support with compulsory reflection, and open AI collaboration.
- Students completed a baseline task, three intervention assignments, and a supervised Week 8 independent no-AI near-transfer task; AI-supported tasks used ChatGPT with GPT-5.3 through the official browser interface.
- Analyses used adjusted linear models with small-sample-corrected class-clustered covariance, t(5) confidence intervals, and exact Webb six-point wild-cluster inference; mechanism analyses were restricted to the 112 AI-exposed students and treated as exploratory, with 5,000 bootstrap intervals for indirect estimates.
Limitations
- Stated: the principal limitation is allocation by only six intact classes, so condition remained partially confounded with class, timetable, and cohort, and wild-cluster inference was fragile.
- Stated: bounded support bundled delegation limits with compulsory reflection, so their independent effects cannot be isolated; Week 8 measured immediate same-course, same-genre near transfer rather than delayed or far transfer; AI/online-learning topics created possible demand characteristics and imperfect content-based blinding; the offloading framework requires independent validation; self-generated text and active drafting time were derived from available records rather than complete keystroke logging; and the sample was Chinese L1 students writing in English at one university.
- Evident from the text: the mediation and moderation analyses are explicitly exploratory and embedded in a quasi-experimental contrast, so they do not establish causal mediation.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
5 Valid · 2 Not valid · 8 Not checked · 123 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“It can also reduce later retention or internal processing when learners cease to encode or practice the offloaded operation”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Literature review > Layer-sensitive cognitive offloading
What the sources say
- Not checked — no full text of the sourcecites Sparrow et al., 2011 source
- Partly supportedcites Grinschgl et al., 2021 sourceExperimental — accuracy not yet measured
The Conclusion reports that offloading diminished subsequent memory for offloaded information, and the Consequences of cognitive offloading section says it can reduce internal processing and elaboration. But Experiment 1 says participants still processed every unit of information; the paper does not establish that they ceased encoding or practicing the offloaded operation. (LLM, conf=0.85)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“deeper idea and reasoning offloading”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Abstract > Results
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Open collaboration predicted higher aggregate offloading after covariate adjustment, a = 0.77, SE = 0.06, p < 0.001, mediator-model R2 = 0.67.”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Results > Exploratory associative decomposition within AI-exposed students
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Coded–self-report convergence correlations were 0.73, 0.70, 0.66, and 0.77.”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Method > Measures and measurement quality
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“deeper idea and reasoning offloading”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Abstract > Results
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
60 more claims
Create a free account to see all 138 claims and every source verdict.
- Probably not a substantial claim. 44
- It may state two separate claims as one. 39
- The quote only partly states this claim. 3
Statistics recomputed
Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.
- Consistentχ²(98) = 129.37reported p = .019 · recomputed p = .019
“χ2(98) = 129.37, p = .019”
- ConsistentF(2, 5) = 0.3reported p = .754 · recomputed p = .753
“F(2, 5) = 0.30, p = .754”
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.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Results > Moderation by self-regulated writing
“Idea and reasoning offloading show the largest negative indirect estimates, while higher self-regulated writing attenuates the negative offloading association.”
§ Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes > Results > Exploratory associative decomposition within AI-exposed students
“| Structural | 0.81 (0.08) | < 0.001 | 0.57 | −0.30 (0.06) | < 0.001 | −0.10 (0.07) | 0.160 | 0.61 | −0.24 | [−0.36, −0.12] | | Idea | 0.92 (0.08) | < 0.001 | 0.62 | −0.21 (0.07) | 0.001 | −0.15 (0.08) | 0.073 | 0.57 | −0.20 | [−0.33, −0.07] |”
Citations
Cites
Cognitive Load During Problem Solving: Effects on Learning
Sweller, J. (1988) · Cognitive Science · doi:10.1016/0364-0213(88)90023-7 · Cited by 11,048 · cited 1×
“Notes, dictionaries, outlines, and feedback can reduce peripheral demands and free resources for more important reasoning, consistent with the principle that reducing extraneous cognitive load can facilitate learning (Sweller, 1988).”
THE ROLE OF TUTORING IN PROBLEM SOLVING *
Wood, D., Bruner, J. S., & Ross, G. (1976) · Journal of Child Psychology and Psychiatry · doi:10.1111/j.1469-7610.1976.tb00381.x · Cited by 8,851 · cited 2×
“Classical scaffolding is contingent, supports learner participation, and transfers responsibility to the learner as competence develops (Wood et al., 1976; van de Pol et al., 2010).”
ChatGPT for good? On opportunities and challenges of large language models for education
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., et al. (2023) · Learning and Individual Differences · doi:10.1016/j.lindif.2023.102274 · Cited by 7,326
The Role of Goal Orientation in Self-Regulated Learning
Pintrich, P. R. (2000) · Handbook of Self-Regulation · doi:10.1016/b978-012109890-2/50043-3 · Cited by 5,173 · cited 1×
“Self-regulated writers set goals, monitor coherence, evaluate feedback, manage effort, and retain authority over revision (Zimmerman and Risemberg, 1997; Pintrich, 2000; Panadero, 2017).”
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja, M., et al. (2023) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2023.102642 · Cited by 4,584
A Review of Self-regulated Learning: Six Models and Four Directions for Research
Panadero, E. (2017) · Frontiers in Psychology · doi:10.3389/fpsyg.2017.00422 · Cited by 2,915 · cited 1×
“Self-regulated writers set goals, monitor coherence, evaluate feedback, manage effort, and retain authority over revision (Zimmerman and Risemberg, 1997; Pintrich, 2000; Panadero, 2017).”
What Is the Impact of ChatGPT on Education? A Rapid Review of the Literature
Lo, C. K. (2023) · Education Sciences · doi:10.3390/educsci13040410 · Cited by 2,024 · cited 1×
“Reviews and conceptual analyses identify substantial opportunities for feedback, personalization, and writing support, alongside concerns about accuracy, overreliance, academic integrity, and assessment validity (Kasneci et al., 2023; Lo, 2023; Dwivedi et al., 2023).”
Scaffolding in Teacher–Student Interaction: A Decade of Research
van de Pol, J., Volman, M., & Beishuizen, J. (2010) · Educational Psychology Review · doi:10.1007/s10648-010-9127-6 · Cited by 1,873
Google Effects on Memory: Cognitive Consequences of Having Information at Our Fingertips
Sparrow, B., Liu, J., & Wegner, D. M. (2011) · Science · doi:10.1126/science.1207745 · Cited by 1,762
Risko, E. F., & Gilbert, S. J. (2016) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2016.07.002 · Cited by 1,502
Research works printed but not linked: 4
- Facione, 1990, Critical Thinking: a Statement of Expert Consensus for Purposes of Educational Assessment and Instruction
- Graham, 2007, Writing Next: Effective Strategies to Improve Writing of Adolescents in Middle and High Schools
- Kellogg, 1996, “A model of working memory in writing,”, The Science of Writing: Theories, Methods, Individual Differences, and Applications, 57
- Perkins, 1992, Transfer of learning, Int. Encyclop. Educ, 2, 6452
Cited by
Ayala, M. G. L., & Osuna-Flores, I. (2026) · Revista Científica Arbitrada de la Fundación MenteClara · doi:10.32351/rca.v11.443
“Chen (2026) found that performance in AI-assisted writing and subsequent no-AI performance may diverge depending on the cognitive layer transferred to the system.”
Influential
Metadata
- Authors
- Chen, X.
- Affiliations
- Jinggangshan University (CN)
- Corresponding author
- Chen, Xinran
- Venue
- Frontiers in Psychology
- Publisher
- Frontiers Media SA
- Published
- 2026-08-28
- Volume
- 17
- Pages
- 1906199-1906199
- ISSN
- 1664-1078 (ISSN-L)
- PMC
- PMC13562006
- OpenAlex
- W7204495850
- Cited by
- 1
- References
- 26
- Citations by year
- 2026: 1
- FWCI
- 2.44
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Psychology (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Social Sciences › Education › Writing and Handwriting Education
- Social Sciences › Psychology › Developmental and Educational Psychology › Innovative Teaching and Learning Methods
- Keywords
- generative AI, cognitive offloading, academic writing, AI-assisted writing, higher-order thinking skills, self-regulated writing, near transfer, writing quality
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: Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes. https://app.cochat.ai/papers/10.3389/fpsyg.2026.1906199
- The paper
- Chen, X. (2026). Layer-sensitive cognitive offloading in generative AI-assisted writing: supported performance and independent no-AI outcomes. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1906199
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.