AI-assisted design synthesis and human creativity in engineering education
Open accessThe DOI matches its Crossref record; 25 of the 25 references it cites were found in a registry or an index; 3 of the 13 sources its claims cite were checked for support.
Abstract
The growing integration of AI into educational and professional settings raises urgent questions about how human creativity evolves when intelligent systems guide, constrain, or accelerate the design process. Generative AI offers structured suggestions and rapid access to ideas, but its role in adopting genuine innovation remains contested. This paper investigates the dynamics of human-AI collaboration in challenge-based design experiments, applying established creativity metrics: fluency, flexibility, originality, and elaboration in order to evaluate outcomes and implications in an engineering education context. Through an exploratory quasi-experimental study, a comparison of AI-assisted and human-only teams was conducted across four dimensions of creative performance: quantity, variety, uniqueness, and quality of design solutions. Findings point to a layered outcome: although AI accelerated idea generation, it also encouraged premature convergence, narrowed exploration, and compromised functional refinement. Human-only teams engaged in more iterative experimentation and produced designs of higher functional quality and greater ideational diversity. Participants' self-perceptions of creativity remained stable across both conditions, highlighting the risk of cognitive offloading, where reliance on AI may reduce genuine creative engagement while masking deficits through inflated confidence. Importantly, cognitive offloading is not directly measured in this study; rather, it is introduced here as a theoretically grounded interpretive explanation that helps contextualize the observed disconnect between performance outcomes and self-perceived creativity. These results bring opportunities and risks. On the one hand, AI can support ideation and broaden access to concepts; on the other, overreliance risks weakening iterative learning and the development of durable creative capacities. The ethical implications are significant, raising questions about accountability and educational integrity when outcomes emerge from human-AI co-creation. The study argues for process-aware and ethically grounded frameworks that balance augmentation with human agency, supporting exploration without eroding the foundations of creative problem-solving. The study consolidates empirical findings with conceptual analysis, advancing the discussion on when and how AI should guide the creative process and providing insights for the broader debate on the future of human-AI collaboration.
Overview
CoChat Summary
## TL;DR
This exploratory quasi-experimental workshop study compared AI-assisted and human-only engineering teams on a catapult design task and found that the AI-assisted teams generally underperformed on creative and functional outcomes. The authors argue that AI may speed ideation but also promote premature convergence and reduce iterative problem-solving, while participants’ self-rated creativity stayed largely unchanged.
## Key findings
- The sample was AI = 7 teams and non-AI = 5 teams; Mann–Whitney U tests found no statistically significant between-group differences at α = 0.05, though effect sizes ranged from r = 0.26–0.56 and favored non-AI teams.
- On design exploration, 60% (3/5) of non-AI teams considered three or more design variants, compared with 0% of AI-assisted teams; this comparison was reported as U = 6, p = 0.073, r = −0.52.
- For perceived uniqueness, 40% of non-AI teams rated their final designs as “very unique” versus 14% (1/7) of AI teams; the paper reports U = 8.5, p = 0.168, r = −0.40.
- Functional performance showed the clearest gap: 100% of non-AI teams’ catapults “worked well,” compared with 29% (2/7) of AI-assisted teams; the paper reports U = 5.0, p = 0.051, r = −0.56.
- Individual self-ratings of creative self-efficacy, divergent-thinking confidence, problem-solving efficacy, and innovative-thinking changed very little in either condition, with Wilcoxon |z| < 1.2, p > 0.20 for all comparisons.
## Methods
- Design: Exploratory, comparative, quasi-experimental study embedded in a 90-minute workshop, using a between-subjects design with AI-assisted and non-AI teams.
- Sample/task: Convenience cohort of volunteers (postgraduate students, early-career lecturers, and industry mentors) self-organized into eight teams of 2–3 people; they completed a 45-minute Catapult Challenge to build a tabletop catapult.
- Analysis: Team outcomes were assessed on Quantity, Variety, Uniqueness, and Quality using survey-based ratings; data were analyzed in R 4.3 with Mann–Whitney U tests for between-group comparisons and Wilcoxon signed-rank tests for pre/post self-ratings, using α = 0.05 and no multiple-comparison correction.
## Limitations
- The paper states that full individual-level randomization was absent because participants self-organized into teams before condition assignment, which creates possible self-selection effects and unobserved confounding; the authors say the findings should not be interpreted as causal.
- The study itself notes that it relies on self-reported creativity measures, which may be biased or miscalibrated, especially in AI-assisted contexts.
- The text also makes evident that the study had a very small sample size and was conducted in a short, single workshop task, which limits statistical power and generalizability.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures




Verification
Claims the paper makes
2 Valid · 0 Not valid · 7 Not checked · 41 Own results and interpretations
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
“Mann–Whitney U tests revealed no statistically significant differences between AI-assisted and non-AI teams at α = 0.05.”
§ AI-assisted design synthesis and human creativity in engineering education > Results > Team-level design outcomes
plain- 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.
- Rests on 2 assumptions not shown here.
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
“Figure 4 provides an integrated perspective through radar visualization, illustrating how non-AI teams consistently outperformed AI-assisted teams across the four creativity dimensions.”
§ AI-assisted design synthesis and human creativity in engineering education > Results > Team-level design outcomes
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.
- Rests on 2 assumptions not shown here.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Non-AI teams proceeded directly to building prototypes, allowing for more iterative testing and improvement cycles.”
§ AI-assisted design synthesis and human creativity in engineering education > Results > Qualitative observations
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.
- Rests on an assumption not shown here.
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
“AI-assisted teams spent substantial time formulating prompts, interpreting ChatGPT responses, and adapting suggested concepts to fit available materials.”
§ AI-assisted design synthesis and human creativity in engineering education > Results > Qualitative observations
plain- 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.
- Rests on an assumption not shown here.
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
“While mechanical variety of designs showed modest differences between conditions (40% vs. 29% achieving distinct principles with U = 11.5, p = 0.373, and r = −0.26)”
§ AI-assisted design synthesis and human creativity in engineering education > Results > Team-level design outcomes
plain- 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.
- Rests on 2 assumptions not shown here.
16 more claims
Create a free account to see all 50 claims and to verify the 2 references still open.
- Probably not a substantial claim. 21
- It may state two separate claims as one. 13
- The quote only partly states this claim. 1
Citations
Cites
Guilford, J. P. (1950) · American Psychologist · doi:10.1037/h0063487 · Cited by 3,588
Social psychology of creativity: A consensual assessment technique.
Amabile, T. M. (1982) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.43.5.997 · Cited by 2,278
Generative artificial intelligence enhances creativity but reduces the diversity of novel content
Doshi, A., & Hauser, O. (2023) · doi:10.2139/ssrn.4535536 · Cited by 1,350
How does generative artificial intelligence impact student creativity?
Habib, S., Vogel, T., Anli, X., & Thorne, E. (2024) · Journal of Creativity · doi:10.1016/j.yjoc.2023.100072 · Cited by 364 · cited 3×
“While AI systems demonstrate remarkable capabilities in combining knowledge in increasingly sophisticated ways, their capacity for true innovation is fundamentally constrained by their reliance on pre-existing data (Habib et al., 2024).”
Computational Creativity: The Final Frontier?
Simon, C., & Wiggins Geraint, A. (2012) · Frontiers in Artificial Intelligence and Applications · doi:10.3233/978-1-61499-098-7-21 · Cited by 350
Artificial Intelligence & Creativity: A Manifesto for Collaboration
Vinchon, F., Lubart, T., Bartolotta, S., Gironnay, V., Botella, M., Bourgeois-Bougrine, S., Burkhardt, J., Bonnardel, N., Corazza, G. E., Glăveanu, V., et al. (2023) · The Journal of Creative Behavior · doi:10.1002/jocb.597 · Cited by 273
Best humans still outperform artificial intelligence in a creative divergent thinking task
Koivisto, M., & Grassini, S. (2023) · Scientific Reports · doi:10.1038/s41598-023-40858-3 · Cited by 247
Artificial muses: Generative artificial intelligence chatbots have risen to human-level creativity
Haase, J., & Hanel, P. H. P. (2023) · Journal of Creativity · doi:10.1016/j.yjoc.2023.100066 · Cited by 233
Artificial intelligence as a tool for creativity
Ivčević, Z., & Grandinetti, M. (2024) · Journal of Creativity · doi:10.1016/j.yjoc.2024.100079 · Cited by 224
Hubert, K. F., Awa, K. N., & Zabelina, D. L. (2024) · Scientific Reports · doi:10.1038/s41598-024-53303-w · Cited by 221
Research works printed but not linked: 4
- Góes, 2023, Pushing GPT’s creativity to its limits: alternative uses and torrance tests
- Seelig, 2017, Creativity rules (reprint)
- Summers-Stay, 2023, Brainstorm, then select: a generative language model improves its creativity score
- Torrance, 1966, Torrance tests of creative thinking—norms technical manual research edition—verbal tests, forms A and B—figural tests, forms A and B
Cited by
İlik, E. B. (2026) · EKEV Akademi Dergisi · doi:10.17753/sosekev.1924403
Wang, C. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1950163
Ramakrishnan, R. (2026) · Design and Artificial Intelligence · doi:10.1016/j.daai.2026.100104 · Cited by 1
Papaneophytou, C. P., & Nicolaou, S. A. (2026) · Trends in Higher Education · doi:10.3390/higheredu5030084
Metadata
- Authors
- Tsakalerou, M., Akhmadi, S., Balgynbayeva, A., & Kumisbek, Y.
- Affiliations
- Nazarbayev University (KZ)
- Corresponding author
- Akhmadi, Saltanat
- Venue
- Frontiers in Artificial Intelligence
- Publisher
- Frontiers Media SA
- Published
- 2026-01-20
- Volume
- 9
- Pages
- 1714523-1714523
- ISSN
- 2624-8212 (ISSN-L)
- PMC
- PMC12864478
- OpenAlex
- W7125008687
- Cited by
- 11
- References
- 22
- Citations by year
- 2026: 11
- FWCI
- 19.92
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Artificial Intelligence (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Nazarbayev University Repository (Nazarbayev University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Nazarbayev University
- Topics
- Physical Sciences › Engineering › Mechanical Engineering › Design Education and Practice
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Creativity in Education and Neuroscience
- Social Sciences › Education › Engineering Education and Pedagogy
- Keywords
- generative AI, human-AI collaboration, creativity, engineering education, design synthesis, cognitive offloading, self-perceived creativity, premature convergence, idea generation, creativity metrics, human agency, AI ethics in education, design 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 2026-10-09, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-09 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
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: AI-assisted design synthesis and human creativity in engineering education. https://app.cochat.ai/papers/10.3389/frai.2026.1714523
- The paper
- Tsakalerou, M., Akhmadi, S., Balgynbayeva, A., & Kumisbek, Y. (2026). AI-assisted design synthesis and human creativity in engineering education. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1714523
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.