pith. sign in

arxiv: 2605.21361 · v1 · pith:GBTGCDSAnew · submitted 2026-05-20 · 💻 cs.HC

Gen-AI-tecture: using generative AI to support architectural students in design tasks

Timo Kapsalis This is my paper

Pith reviewed 2026-05-21 03:33 UTC · model grok-4.3

classification 💻 cs.HC
keywords generative AIarchitectural educationcreativityinclusivitydesign tasksAI skillspedagogyfocus groups
0
0 comments X

The pith

Generative AI tools enhance creative fluency and inclusivity for architecture students in design tasks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper tests a custom generative AI tool inside architecture courses through focus groups to measure effects on creativity, inclusivity across learner types, and AI skill building. It aims to show that such tools can support students in producing more fluent designs while opening participation to varied profiles and preparing them for technology-heavy careers. If the results hold, architecture programs could adopt similar local tools to make design education more accessible and aligned with how students learn today. The work supplies practical steps for adding AI workflows that fit constructivist ideas of students creating their own understanding.

Core claim

The Gen-AI-tecture project embeds a locally executed, discipline-specific generative AI tool into a mixed-methods focus-group design and finds enhanced creative fluency in design-thinking processes and outcomes, broadened participation across diverse learner profiles, and strengthened confidence in AI-supported design processes. This supplies evidence-based guidance for integrating generative-AI workflows into architectural pedagogy and shows how the tools can operationalise constructivist principles of learner-led meaning-making while advancing inclusive educational practices.

What carries the argument

The Gen-AI-tecture tool, a locally executed discipline-specific generative AI system integrated into focus-group evaluations to assess impacts on creativity, inclusivity, and AI-handling skills.

If this is right

  • Students gain enhanced creative fluency in their design processes and final outcomes.
  • Participation widens to include learners with different backgrounds and profiles.
  • Students develop greater confidence and practical skills in using AI for design work.
  • Architecture programs receive concrete guidance for adding generative AI to teaching methods.
  • Learning practices become more inclusive, flexible, and aligned with contemporary student needs.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same local-tool approach could be adapted and tested in related fields such as product design or landscape architecture.
  • Longer-term tracking of graduates might reveal whether early AI exposure improves job performance or career progression.
  • Schools could develop rules for ethical AI use in student assessments to prevent over-dependence on the tools.
  • Hybrid models that combine the AI tool with traditional studio methods may balance creative gains against risks of reduced original thinking.

Load-bearing premise

The mixed-methods focus-group design with its self-reported data and limited participant group accurately isolates the tool's effects on creativity and inclusivity without major influence from reporting bias or the act of taking part in the study.

What would settle it

A larger randomized trial that scores design projects blindly for creativity and finds no measurable gain for students using the generative AI tool compared with a control group working without it.

read the original abstract

The "Gen-AI-tecture" project embeds a locally executed, discipline-specific tool into a mixed-methods focus-group design, structured around three research objectives: (a) to evaluate how generative AI tools impact students' creativity in design-thinking processes and outcomes, (b) to assess whether these tools enhance inclusivity in learning processes, and (c) to examine how they develop students' AI-handling skills with a view to boosting future employability. Findings indicate enhanced creative fluency, broadened participation across diverse learner profiles, and strengthened confidence in AI-supported design processes. The study contributes evidence-based guidance for integrating generative-AI workflows into architectural pedagogy, demonstrating how such tools can operationalise constructivist principles of learner-led meaning-making, support connectivist understandings of learning as participation in human-AI networks, and advance universal learning theories by promoting more inclusive, flexible and accessible educational practices for contemporary learners.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript presents the Gen-AI-tecture project, which integrates a locally executed, discipline-specific generative AI tool into architectural education through a mixed-methods focus-group study. It addresses three objectives: evaluating impacts on students' creativity in design-thinking processes, assessing enhancements to inclusivity in learning, and examining development of AI-handling skills for employability. The reported findings indicate enhanced creative fluency, broadened participation across diverse learner profiles, and strengthened confidence in AI-supported design processes, while linking these to constructivist, connectivist, and universal learning theories to provide guidance for AI integration in architectural pedagogy.

Significance. If the empirical claims hold after addressing methodological gaps, the work could supply timely, evidence-based recommendations for incorporating generative AI workflows into design education. It would strengthen the case for tools that operationalize learner-led meaning-making and human-AI participation networks, with potential relevance to broader HCI and educational technology audiences seeking inclusive practices.

major comments (2)
  1. [Abstract / Research Design] Abstract and study design description: The central claims of enhanced creative fluency, broadened participation, and strengthened confidence rest on focus-group observations, yet no sample size, recruitment details, specific measures (e.g., creativity rubrics or scales), control conditions, or analysis procedures are supplied. This prevents verification that observed outcomes can be attributed to the Gen-AI tool rather than self-report bias, social desirability, or participation effects, directly undermining the isolation assumption required for the findings.
  2. [Findings / Discussion] Findings and discussion sections: The manuscript reports qualitative themes supporting inclusivity and confidence gains but provides no pre/post objective metrics, blinded expert ratings of design artifacts, or quantitative scales to corroborate self-reports. Without these, the attribution of changes to the locally executed tool versus Hawthorne-like or group-dynamic confounds remains unsecured and load-bearing for the pedagogical contribution claim.
minor comments (2)
  1. [Methods] Clarify the exact technical implementation of the 'locally executed' tool (e.g., model, interface constraints) to allow replication by other architectural educators.
  2. [Title / Objectives] Ensure consistent terminology between 'design-thinking processes' in the objectives and 'design tasks' in the title to avoid minor reader confusion.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive feedback on our manuscript. We value the recognition of the potential contribution of the Gen-AI-tecture project to architectural pedagogy and HCI. We address each major comment below, providing clarifications on our qualitative focus-group design while committing to revisions that improve transparency without altering the study's exploratory scope.

read point-by-point responses

  1. Referee: [Abstract / Research Design] Abstract and study design description: The central claims of enhanced creative fluency, broadened participation, and strengthened confidence rest on focus-group observations, yet no sample size, recruitment details, specific measures (e.g., creativity rubrics or scales), control conditions, or analysis procedures are supplied. This prevents verification that observed outcomes can be attributed to the Gen-AI tool rather than self-report bias, social desirability, or participation effects, directly undermining the isolation assumption required for the findings.

    Authors: We agree that greater methodological transparency is needed. Our study is an exploratory mixed-methods focus-group investigation rather than a controlled experiment, so it does not include control conditions or claim strict causal isolation. In the revised manuscript we will expand the methods section to report the exact sample size and participant demographics, recruitment process, discussion prompts used to elicit themes on creativity and inclusivity, and the thematic analysis procedures (including coding steps and inter-coder reliability checks if applicable). These additions will allow readers to better assess the grounding of the reported themes. revision: yes

  2. Referee: [Findings / Discussion] Findings and discussion sections: The manuscript reports qualitative themes supporting inclusivity and confidence gains but provides no pre/post objective metrics, blinded expert ratings of design artifacts, or quantitative scales to corroborate self-reports. Without these, the attribution of changes to the locally executed tool versus Hawthorne-like or group-dynamic confounds remains unsecured and load-bearing for the pedagogical contribution claim.

    Authors: The study deliberately employs a qualitative focus-group approach to capture in-depth student perceptions and experiences in a naturalistic educational setting, consistent with constructivist and connectivist frameworks. Pre/post quantitative metrics and blinded expert ratings of artifacts were not part of the original design. We will revise the discussion to explicitly acknowledge potential confounds such as social desirability and group dynamics, clarify that findings are interpretive and preliminary rather than definitive causal claims, and strengthen the positioning of the work as evidence-based guidance for AI integration rather than a controlled efficacy trial. revision: partial

Circularity Check

0 steps flagged

No circularity: empirical claims grounded in focus-group observations

full rationale

The paper presents findings from a mixed-methods focus-group study evaluating a locally executed Gen-AI tool in architectural education. Central claims of enhanced creative fluency, broadened participation, and strengthened confidence are reported as direct outcomes of the described research design and participant responses rather than any derivation, equation, or self-referential construct that reduces to the paper's own inputs. No mathematical models, fitted parameters presented as predictions, uniqueness theorems, or load-bearing self-citations appear in the provided text. The study is self-contained as an empirical report with independent grounding in qualitative data collection.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

As an empirical education study, the central claim rests primarily on standard qualitative research assumptions rather than mathematical derivations, fitted parameters, or new postulated entities.

axioms (1)
  • domain assumption Mixed-methods focus groups can reliably evaluate impacts on creativity, inclusivity, and skill development in educational settings
    Invoked through the study design and objectives to support conclusions drawn from qualitative data.

pith-pipeline@v0.9.0 · 5678 in / 1336 out tokens · 56992 ms · 2026-05-21T03:33:57.222433+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

10 extracted references · 10 canonical work pages · 1 internal anchor

  1. [1]

    and Ashmeel, R

    Alamasi, R., Asfour, O.S. and Ashmeel, R. (2026) ‘The impact of generative AI on architectural design education: insights from hands-on experience with architecture students’ , Ain Shams Engineering Journal, 17(1), 103879. https://doi.org/10.1016/j.asej.2025.103879 Aly, A., Elazazy, A. and Sharaf, N. (2025) ‘Integrating generative AI in architectural educ...

    work page doi:10.1016/j.asej.2025.103879 2026

  2. [2]

    and Owenz, M.B

    https://doi.org/10.3390/educsci15091160 Hyatt, S.E. and Owenz, M.B. (2024) ‘Using Universal Design for Learning and Artificial Intelligence to Support Students with Disabilities’ , College Teaching, pp. 1–8 Jin, S., Tu, H., Li, J., Fang, Y ., Qu, Z., Xu, F . et al. (2024) ‘Enhancing architectural education through artificial intelligence: a case study of an...

    work page doi:10.3390/educsci15091160 2024

  3. [3]

    InteriorNet: Mega-scale Multi-sensor Photo-realistic Indoor Scenes Dataset

    https://doi.org/10.3390/buildings14061613 Kahn, K. & Winters, N. (2021) ‘Constructionism and AI: A history and possible futures’ , British Journal of Educational Technology, 52(3), pp. 1130–1142. doi:10.1111/bjet.13088. Kotsis, K.T. (2024) ‘Integrating ChatGPT into the inquiry-based science curriculum for primary education’ , European Journal of Education...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3390/buildings14061613 2021

  4. [4]

    https://lumivero.com/products/nvivo/ Luo, J., Zhao, M

    [Software]. https://lumivero.com/products/nvivo/ Luo, J., Zhao, M. and Zhang, H., (2025). Low-rank adaptation for parameter-efficient fine-tuning in composed image retrieval. In: Proceedings of the 2025 International Conference on Multimedia Retrieval (ICMR ’25). Association for Computing Machinery, pp. 963–971. doi:10.1145/3731715.3733377. Melo-L ó p ez ,...

    work page doi:10.1145/3731715.3733377 2025

  5. [5]

    and Gordon, D

    https://doi.org/10.3390/educsci15050539 Meyer, A., Rose, D.H. and Gordon, D. (2014) Universal Design for Learning: Theory and Practice. Wakefield, MA: CAST Moundridou, M., Matzakos, N. & Doukakis, S. (2024) ‘Generative AI tools as educators’ assistants: Designing and implementing inquiry-based lesson plans’ , Computers and Education: Artificial Intelligence...

    work page doi:10.3390/educsci15050539 2014

  6. [6]

    (2025) ‘The role of generative AI in shifting the paradigm of constructivist learning models’ , preprint available via ResearchGate

    ScienceDirect+1 Owen, A. (2025) ‘The role of generative AI in shifting the paradigm of constructivist learning models’ , preprint available via ResearchGate. Paananen, V ., Oppenlaender, J. and Visuri, A. (2024) ‘Using text-to-image generation for architectural design ideation’ , International Journal of Architectural Computing, 22(3), pp. 458–474. https:...

    work page doi:10.1177/14780771231222783 2025

  7. [7]

    Saborío-Taylor, S

    London: Royal Institute of British Architects. Saborío-Taylor, S. and Rojas-Ramírez, F . (2024) ‘Universal Design for Learning and Artificial Intelligence in the Digital Era: Fostering Inclusion and Autonomous Learning’ , International Journal of Professional Development, Learners and Learning, 6(2), ep2408. Sánchez Muñoz, J.A. et al. (2025) ‘GenAI as a co...

    work page doi:10.3389/feduc.2025.1597249 2024

  8. [8]

    more knowledgeable other

    Tran, M. et al. (2025) ‘Generative artificial intelligence: The “more knowledgeable other” in a social constructivist framework of medical education’ , npj Digital Medicine, 8,

    work page 2025

  9. [9]

    Varsik, S

    doi:10.1038/s41746-025-01823-8. Varsik, S. and Vosberg, L. (2024) The Potential Impact of Artificial Intelligence on Equity and Inclusion in Education. OECD Artificial Intelligence Papers, No

    work page doi:10.1038/s41746-025-01823-8 2024

  10. [10]

    Veytia Bucheli, M.G., Gómez-Galán, J., Cáceres Mesa, M.L

    Paris: OECD Publishing. Veytia Bucheli, M.G., Gómez-Galán, J., Cáceres Mesa, M.L. (2024) Digital technologies as enablers of universal design for learning: higher education students' perceptions in the context of SDG4. Discov Sustain 5, 473 (2024). https://doi.org/10.1007/s43621-024- 00699-0 Yu, G. (2025) ‘Toward understanding the role of generative AI in...

    work page doi:10.1007/s43621-024- 2024