REVIEW 4 major objections 5 minor 40 references
The Cloud Weaving Model for AI development
T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash
Pith's one-line read The paper argues that AI co-creation with marginalized communities fails less from bad execution than from treating AI as a deliverable instead of a relational practice that is never finished.
desk verdict A coherent, honestly scoped metaphor for community-centered AI whose claimed payoff—surfacing neglected dimensions—is asserted rather than demonstrated. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the five-element metaphor and its associated weaving processes: anchoring (building trust with a first community), linking (bridging to further communities against the weather), bending at the center (letting the problem definition shift as new spokes are added), spiralling (approaching but never reaching the center of the problem), and encountering other spiders (coordinating or conflicting with other AI systems). The metaphor carries the argument by supplying an interpretive vocabulary in which the paper re-describes the CommuniCity patterns; its five elements map directly onto dimensions that the authors argue current software-engineering and pilot-management paradigms omit, such as long-term trust, collective rather than individual agency, and systemic power as an active force.
What would settle it
Give two teams the same set of pilot logs and interview transcripts from a co-creation project; ask one team to analyze them with the Cloud Weaving Model and the other with a standard participatory-design or human-centered design checklist. If the second team produces the same diagnoses and recommended actions without needing the five elements, the model's contribution is purely terminological. A weaker test: if no practitioner, after training on the model, reports a change in what they notice or decide, the claimed foregrounding of conceptual limitations is not observable.
Extended reading notes
Core claim
The paper's central claim is that socially aware AI is best conceived as a practice rather than a product: an ongoing act of weaving that connects communities through information and care, and that is never finished because communities shift and systemic power (weather) constantly damages the web. The five elements are not decorative; each names a dimension that mainstream piloting frameworks neglect. Clouds are not individuals but overlapping, flowing collectives that include humans, non-humans, and natural resources; the spider is an AI that is always subject to the clouds it connects and cannot stand outside them; the thread carries not only data but trust, oral knowledge, and embodied experience; the web is the structural support that provides systemic change and individual relief while its center, the problem, wobbles as spokes are added; weather is the inescapable power structure that determines whose perspectives count. The paper applies this vocabulary to five patterns from the CommuniCity pilots and reads each as a disturbance of weaving, whether threads never become durable, webs stay too fragile, sustenance is prioritized over clouds, webs break and must be rebuilt, or no technical web is possible at all. The conclusion is that evaluation and design should center the perseverance of weaving rather than one-off delivery.
Load-bearing premise
The framework rests on the assumption that mapping clouds to communities, spider to AI, thread to information, spiderweb to socio-technical solution, and weather to systemic power is a valid and productive analogy rather than an arbitrary renaming of concepts participatory design already has.
Editorial extensions
If this is right
- Pilots should treat trust-building with the first community as a generation-spanning commitment, not a pre-phase to be rushed.
- Success metrics for co-created AI should include relational outcomes: continued engagement, mutual learning, and web-repair capacity, alongside or instead of delivered outputs.
- Evaluation cannot assume a stable problem: the center of the web shifts as new spokes are added, so problem definitions must be revisited after each community connection.
- Systemic power, the weather, must be addressed directly; platform designs that empower individuals while ignoring structural barriers will produce fragile webs.
- Non-technical outcomes are legitimate: where weather is too extreme, the right conclusion may be that no technical web can be woven.
Reading between the lines
- Editorial inference: the model implies that fixed-duration grant structures are structurally mismatched to community AI work; a testable extension would compare pilot outcomes across funding horizons while holding the project type fixed.
- Editorial inference: the web-of-webs idea suggests that AI systems should be assessed for how they interact with other AI systems and existing community support structures, not only for their direct effects.
- Editorial inference: the weather element could be operationalized by coding, for each co-creation pilot, which systemic forces (funding rules, language regimes, legal status) are named and whether the design responds to them.
- Editorial inference: if AI is a practice, documentation of failed or partial weaving becomes a public good; a concrete extension is to require projects to publish relationship logs and web-repair notes as part of responsible AI reporting.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes the Cloud Weaving Model, a conceptual metaphor for AI development that maps clouds to communities and stakeholders, spiders to AI systems, threads to information/trust/care, spiderwebs to socio-technical support systems, and weather to systemic power structures. It derives five process concepts (anchoring, linking, bending at the center, spiralling, encountering other spiders) and uses the element vocabulary to interpret five patterns from the CommuniCity project: a VR mental-health pilot, community engagement platforms (girls' sports, elderly loneliness), a refugee translation app, pilots with knowledge/language gaps, and a pilot concluding that no technical solution was possible. The authors argue that AI should be seen as a practice rather than a product and position the model as an interpretative tool for critical reflection rather than as a practical methodology. The empirical basis is described in Section 1 and footnote 1 as dedicated observations of CommuniCity open calls and analysis of publicly available project documents.
Significance. If accepted as an interpretive lens, the model has genuine value for HCI and community-centered AI: it supplies a relational vocabulary that treats AI as an ongoing practice embedded in communities, foregrounds systemic power through the 'weather' element, and draws on non-Western epistemologies in a way that most software-engineering-oriented piloting frameworks do not. The authors are honest about scope: Section 5 explicitly states that the model is primarily an interpretative tool for critical reflection and that practical translation is left to future research, and the paper appropriately acknowledges the diversity of Indigenous and Eastern traditions. The five processes are evocative and internally coherent. However, the contribution is currently more a metaphor proposal than a demonstrated framework: the claimed ability to highlight neglected dimensions is not compared against existing participatory design or design justice vocabularies, and the illustrative application relies on post hoc, under-specified observation. The significance is accordingly conditional on the additional grounding that a revision should provide.
major comments (4)
- [Section 4, Patterns 1-5] The paper's central claim, stated in the abstract and in Section 5, that the Cloud Weaving Model 'highlights neglected yet relevant dimensions' is not demonstrated, because the Section 4 analysis never compares the model's output with existing community-centered frameworks such as participatory design, value-sensitive design, or design justice. The five patterns are fully expressible without the new vocabulary (six-month funding cycles without exit strategies, language and trust barriers, structural barriers to girls' sports, knowledge gaps between developers and unemployed people, and cases requiring non-technical solutions). To support the 'neglected dimensions' claim, the authors need to show, on the same pilots, what the Cloud Weaving vocabulary brings out that these existing vocabularies do not.
- [Footnote 1 and Section 4 intro] The empirical basis of the analysis is under-specified. The paper mentions dedicated observations of two rounds of open calls and analysis of public documents, but it provides no observation protocol, no data excerpts, no coding or analysis method, and no criteria for selecting the five patterns. Since the model was iteratively built from the CommuniCity experience (Section 5 states it 'originated being grounded in a very concrete localised perspective') and the same project is then used to illustrate the model, the fit between model and examples is at risk of being a projection of the model's own categories. I would ask the authors to either supply a traceable qualitative analysis or explicitly present Section 4 as an illustrative re-description rather than as evidence for the model's utility.
- [Section 3.3] The five processes (anchoring, linking, bending at the center, spiralling, encountering other spiders) are introduced as part of the model but are not used in the Section 4 application, which only draws on the five static elements (clouds, spider, thread, web, weather). Either the processes need to be operationalized and applied to the pilots, or the paper should explicitly say that the process layer is a speculative proposal; as written, a substantial part of the model is left unexercised, which weakens the claim that the model provides a usable analytic lens.
- [Section 3.2] The element mapping is highly elastic: 'clouds' cover people, communities, natural resources, pets, and social media interlocutors, and 'weather' covers systemic power, economic scarcity, gender norms, and literal storms. No criteria are given for deciding when a phenomenon belongs to one element rather than another. Without such criteria, any observation can be redescribed in the model's terms, which makes it difficult to assess the claim that the model brings genuinely neglected dimensions to the foreground. The authors should specify constraints or at least discuss the model's boundary conditions.
minor comments (5)
- [Section 3.3] The phrase 'Based on the prevous five elements' contains a typo; 'prevous' should be 'previous'. Section 5 also contains 'simultaneusly', which should be 'simultaneously'.
- [Section 5] The sentence 'The model at this stage serves primarily as a interpretative tool' uses the wrong article; it should be 'an interpretative tool'.
- [Figure 1] Figure 1 illustrates the five processes but is only mentioned once in the text; a detailed caption or per-process references in Section 3.3 would help readers connect each subfigure to the corresponding process.
- [Section 4.5] The phrase 'the weather conditions were too extreme' uses 'weather' in a literal sense immediately after the model's figurative definition; clarify that the phrase refers to the model's systemic 'weather' rather than meteorology.
- [Section 5] The paper says the model will be mapped to guidelines for ethical AI design currently being developed within CommuniCity, but the cited reference [20] is a Tethics paper on design justice rather than those guidelines; provide a direct citation if such a document exists.
Circularity Check
No significant circularity: the model is presented as an explicitly interpretive vocabulary, and Section 4's application is illustrative rather than predictive or confirmatory.
full rationale
The paper makes no empirical prediction and performs no parameter fitting. The Cloud Weaving Model is introduced in Section 3 as a stipulated interpretive vocabulary, and Section 4 applies it to CommuniCity patterns explicitly as illustration, not validation: 'These patterns were selected here not to be exhaustive in describing the project, but rather to exemplify the value of changing paradigm while self-reflecting over a such complex applied domain.' Section 5 explicitly limits the model to 'primarily a interpretative tool, meant for critical reflection' and leaves practical translation to future research. The self-citations ([16], [18], [20]) are to methodological, conceptual, or future-work sources and are not load-bearing: the argument does not depend on their correctness as theorems or empirical results. The concern that applying the model to the same project that inspired it limits validation is acknowledged by the authors themselves ('this model originated being grounded in a very concrete localised perspective'), and it is a limitation of evidence, not a circular derivation. No equation, definition, or citation chain reduces the paper's claims to its own inputs. Although the metaphor re-describes familiar participatory-design concerns such as trust, long-term commitment, and systemic power, the paper presents this re-description as its intended contribution ('more accessible vocabularies') rather than as an empirical discovery, which does not constitute circularity under the stated criteria.
Assumptions & free parameters
assumptions (4)
- domain assumption It is legitimate and productive to transfer Buddhist, Zen, and Indigenous concepts of clouds and weaving into AI development.
- domain assumption AI systems perpetuate structural oppression unless designers actively counter systemic weather.
- domain assumption Reflection and introspection by the authors are sufficient to construct and apply a conceptual framework.
- domain assumption The selected CommuniCity pilots and public documents are adequate to support the five observed patterns.
invented entities (5)
-
Clouds (communities and stakeholders as dynamic, overlapping entities)
-
Spider (the AI system as an agent that weaves connections)
-
Thread (information, trust, and care connecting stakeholders)
-
Spiderweb (the socio-technical support system under construction)
-
Weather (systemic power structures and social conditions)
Cite this review
Pith. "Pith review of The Cloud Weaving Model for AI development." pith.science (2026). https://pith.science/paper/OWKGJSC2
@misc{pith2026250417823,
author = {Pith},
title = {Pith review of: The Cloud Weaving Model for AI development},
year = {2026},
howpublished = {\url{https://pith.science/paper/OWKGJSC2}},
note = {Machine review of arXiv:2504.17823}
}
read the original abstract
While analysing challenges in pilot projects developing AI with marginalized communities, we found it difficult to express them within commonly used paradigms. We therefore constructed an alternative conceptual framework to ground AI development in the social fabric -- the Cloud Weaving Model -- inspired (amongst others) by indigenous knowledge, motifs from nature, and Eastern traditions. This paper introduces and elaborates on the fundamental elements of the model (clouds, spiders, threads, spiderwebs, and weather) and their interpretation in an AI context. The framework is then applied to comprehend patterns observed in co-creation pilots approaching marginalized communities, highlighting neglected yet relevant dimensions for responsible AI development.
Figures
Reference graph
Works this paper leans on
-
[1]
Ali Alkhatib. 2021. To Live in Their Utopia: Why Algorithmic Systems Create Absurd Outcomes. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems (Yokohama, Japan) (CHI ’21). Association for Computing Machinery, New York, NY, USA, Article 95, 9 pages. https://doi.org/10.1145/ 3411764.3445740
arXiv 2021
- [2]
-
[3]
Francis Bacon. 1620. Instauratio Magna (Novum Organum) . Number Book 1
-
[4]
Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell
Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?. In FAccT ’21: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (Virtual Event, Canada) (FAccT ’21). Association for Computing Machinery, New York, NY, USA, 610–6...
arXiv 2021
-
[5]
Ruha Benjamin. 2022. Viral Justice: How We Grow the World We Want . Princeton University Press, Princeton. 392 pages
work page 2022
-
[6]
Cynthia L. Bennett and Daniela K. Rosner. 2019. The Promise of Empathy: Design, Disability, and Knowing the "Other". In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (Glasgow, Scotland Uk) (CHI ’19) . Association for Computing Machinery, New York, NY, USA, 1–13. https://doi.org/10.1145/3290605.3300528 5A similar situation wa...
arXiv 2019
-
[7]
P. Besserman and M.B. Steger. 1991. Crazy Clouds: Zen Radicals, Rebels, and Reformers. Shambhala, Boston, London
work page 1991
-
[8]
Joy Buolamwini and Timnit Gebru. 2018. Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. In Proceedings of the 1st Confer- ence on Fairness, Accountability and Transparency (Proceedings of Machine Learn- ing Research, Vol. 81), Sorelle A. Friedler and Christo Wilson (Eds.). PMLR, New York, NY, USA, 77–91. https://proc...
work page 2018
Show all 40 references
-
[9]
Daniel Casasanto. 2014. Experiential Origins of Mental Metaphors: Language, Culture, and the Body . American Psychological Association, Washington, DC, US, 249–268. http://www.jstor.org/stable/j.ctv1chs139.15
2014
-
[10]
Chia-chen Chang, Zhi Yun Lim, Danielle A Klomp, Yusoff Norma-Rashid, and Daiqin Li. 2018. Aggressive spiders make the wrong decision in a difficult task. Behavioral Ecology 29, 4 (2018), 848–854. https://doi.org/10.1093/beheco/ary066
2018 doi
-
[11]
CommuniCity. 2022. CommuniCity Project Website. CommuniCity. Retrieved December 3, 2024 from https://communicity-project.eu/
2022
-
[12]
M. Cooley. 2015. Architect Or Bee?: The Human Price of Technology . Spokesman, Nottingham
2015
-
[13]
Harry F. Dahms. 2007. Confronting the Dynamic Nature of Modern Social Life. Soundings: An Interdisciplinary Journal, 90, 3/4 (2007), 191–205. http: //www.jstor.org/stable/41179155
2007
-
[14]
Dogen. 2004. The Treasury of the True Dharma Eye: ZEN Master Dogen’s Shobo Genzo. Sotoshu Shumucho. translated by Carl Bielefeldt
2004
-
[15]
Thomas Hylland Eriksen. 2015. Small Places, Large Issues - Fourth Edition: An Introduction to Social and Cultural Anthropology . Pluto Press, London, UK. http: //www.jstor.org/stable/j.ctt183p184
2015
-
[16]
Sennay Ghebreab. 2023. Infinity, kaleidoscopy and society: On reflections of AI . University of Amsterdam. https://dare.uva.nl/search?identifier=9b7e06c5-6ad6- 40db-8857-8a3a2d637e7d
2023
-
[17]
Raymond W. Gibbs. 2014. Conceptual Metaphor in Thought and Social Action . American Psychological Association, Washington DC, US, 17–40. http://www. jstor.org/stable/j.ctv1chs139.5
2014
-
[18]
Geoff Gordon, Bernhard Rieder, and Giovanni Sileno. 2022. On mapping values in AI Governance. Computer Law & Security Review 46 (2022), 105712. https: //doi.org/10.1016/j.clsr.2022.105712
2022
-
[19]
2014.The radical Imagination : social movement research in the age of austerity
Max Haiven and Alex Khasnabish. 2014.The radical Imagination : social movement research in the age of austerity . Fernwood Publishing, Halifax
2014
-
[20]
Aida Kalender, Giovanni Sileno, and Sennay Ghebreab. 2025. Toward Just Smart Cities: Community-based Arts Organisations as Partners in Design Justice. In Proceedings of the Conference on Technology Ethics 2024 (Tethics 2024),Tampere, Finland, November 6-7, 2024. (CEUR Workshop...
2025
-
[21]
Klinkman, and Mark S
Elizabeth Kaziunas, Michael S. Klinkman, and Mark S. Ackerman. 2019. Precarious Interventions: Designing for Ecologies of Care.Proc. ACM Hum.-Comput. Interact. 3, CSCW, Article 113 (2019), 27 pages. https://doi.org/10.1145/3359215
2019 doi
-
[22]
Khalaf, Abdullah A
Abderrahman M. Khalaf, Abdullah A. Alubied, Ahmed M. Khalaf, and Abdallah A. Rifaey. 2023. The Impact of Social Media on the Mental Health of Adolescents and Young Adults: A Systematic Review. Cureus 15, 8 (Aug. 2023). https: //doi.org/10.7759/cureus.42990
2023 doi
-
[23]
Jason Edward Lewis. 2020. Indigenous Protocol and Artificial Intelligence Position Paper. The Initiative for Indigenous Futures and the Canadian Institute for Ad- vanced Research (CIFAR), Honolulu, Hawaií. https://doi.org/10.11573/spectrum. library.concordia.ca.00986506
2020 doi
-
[24]
Audre Lorde. 1984. The Master’s Tools Will Never Dismantle the Master’s House. Sister Outsider: Essays and Speeches (1984)
1984
-
[25]
Marx and F
K. Marx and F. Engels. 1954. Capital. Capital, Vol. 1. Progress Publishers, Moscow, USSR
1954
-
[26]
Ziad Obermeyer, Brian Powers, Christine Vogeli, and Sendhil Mullainathan. 2019. Dissecting racial bias in an algorithm used to manage the health of populations. Science 366, 6464 (2019), 447–453. https://doi.org/10.1126/science.aax2342
2019 doi
-
[27]
Encyclopedia of Korean Culture. n.a.. Clouds. Encyclopedia of Korean Culture. Retrieved December 3, 2024 from https://encykorea.aks.ac.kr/Article/E0005744
2024
-
[28]
1966–1988
Plotinus. 1966–1988. Enneads Vol. I-VI. Harvard University Press, Cambridge, US. translated by A.H. Armstrong, Loeb Classical Library
1966
-
[29]
Mario Pérez-Montoro and Lluís Codina. 2017. Chapter 5 - The Essentials of Search Engine Optimization. In Navigation Design and SEO for Content-Intensive Websites, Mario Pérez-Montoro and Lluís Codina (Eds.). Chandos Publishing, Cambridge, 109–124. https://doi.org/10.1016/B978-...
2017 doi
-
[30]
Shunryu Suzuki Roshi. 2016. The Mind Itself is Buddha . Tsony. Retrieved December 3, 2024 from https://tsony.com/wp-content/uploads/2016/01/Suzuki- Roshi-The-Mind-Itself-is-Buddha.pdf
2016
-
[31]
Gary Snyder. 1983. Axe handles : poems . North Point Press, San Francisco
1983
-
[32]
Sharifa Sultana, François Guimbretière, Phoebe Sengers, and Nicola Dell. 2018. Design Within a Patriarchal Society: Opportunities and Challenges in Design- ing for Rural Women in Bangladesh. In CHI ’18: Proceedings of the 2018 CHI Conference on Human Factors in Computing Syste...
2018
-
[33]
Saurabh Vaidya, Prashant Ambad, and Santosh Bhosle. 2018. Industry 4.0 – A Glimpse. Procedia Manufacturing 20 (2018), 233–238. https://doi.org/10.1016/j. promfg.2018.02.034
2018 doi
-
[34]
Van der Post
L. Van der Post. 1958. The Lost World of the Kalahari . Morrow, University of California. https://books.google.nl/books?id=8HsNAQAAIAAJ
1958
-
[35]
Vinay Reddy Venumuddala, Amit Prakash, and Bidisha Chaudhuri. 2024. Gov- erning Smart City IoT Interventions: A Complex Adaptive Systems Perspec- tive. Digit. Gov.: Res. Pract. 5, 3, Article 23 (Sept. 2024), 24 pages. https: //doi.org/10.1145/3660644
2024 doi
-
[36]
Vonnegut
M. Vonnegut. 2002. The Eden express: a memoir of insanity . Seven Stories Press, New York
2002
-
[37]
Yi Wang, Hai-Shu Ma, Jing-Hui Yang, and Ke-Sheng Wang. 2017. Industry 4.0: a way from mass customization to mass personalization production. Advances in Manufacturing 5, 4 (01 Dec 2017), 311–320. https://doi.org/10.1007/s40436-017- 0204-7
2017 doi
-
[38]
Peter Wright and John McCarthy. 2008. Empathy and Experience in HCI. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Florence, Italy) (CHI ’08). Association for Computing Machinery, New York, NY, USA, 637–646. https://doi.org/10.1145/1357054.1357156
2008
-
[39]
Chunchen Xu and Xiao Ge. 2024. AI as a Child of Mother Earth: Regrounding Human-AI Interaction in Ecological Thinking. In CHI EA ’24: Extended Abstracts of the CHI Conference on Human Factors in Computing System (Honolulu, HI, USA) (CHI EA ’24). Association for Computing Machi...
2024
-
[40]
Shoshana Zuboff. 2018. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (1st ed.). Profile Books, London. Received 05 December 2024
2018
Reviewed August 16, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.