pith. sign in

arxiv: 2606.29227 · v1 · pith:V4Q47DRWnew · submitted 2026-06-28 · 💰 econ.GN · cs.CY· q-fin.EC

The Human-Machine Knowledge Spiral

Aaron Chatterji , Daniel Rock , Eduard Talamas This is my paper

Pith reviewed 2026-06-30 02:40 UTC · model grok-4.3

classification 💰 econ.GN cs.CYq-fin.EC
keywords knowledge managementtacit knowledgeartificial intelligenceinnovationorganizational learningNonaka modelmachine knowledgeknowledge spiral
0
0 comments X

The pith

AI introduces tacit machine knowledge into Nonaka's conversion processes, yet the company's role in building shared context for knowledge spirals remains unchanged.

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

The paper extends Nonaka's model of innovation as a continuous conversion between tacit and explicit knowledge. It identifies AI outputs as a new category called tacit machine knowledge that participates in the same back-and-forth. The central task for firms stays the creation of organizational contexts where human and machine knowledge interact, convert, and generate further cycles of innovation. A reader would care because this framing treats AI as an addition to existing dynamics rather than a replacement for them. The argument keeps the knowledge-creating company as the unit that turns these interactions into sustained advantage.

Core claim

Nonaka described innovation as arising from the spiral of socialization, externalization, combination, and internalization between tacit and explicit knowledge. The paper claims that artificial intelligence supplies tacit machine knowledge that enters this same spiral, converting with human knowledge to produce organizational knowledge and renewed innovation. The knowledge-creating company therefore retains its core function of establishing the shared context that enables these conversions across human and machine forms.

What carries the argument

The human-machine knowledge spiral, an extension of Nonaka's SECI processes that incorporates tacit machine knowledge from AI as a participant in the same conversion cycle.

If this is right

  • Firms must design shared contexts that allow tacit machine knowledge to interact with human tacit knowledge.
  • AI outputs become part of the explicit knowledge base only after passing through the same conversion steps as human knowledge.
  • Leadership focus shifts to enabling continuous cycles in which machine knowledge feeds human innovation and human knowledge refines machine outputs.
  • The knowledge-creating company remains the locus where these mixed knowledges combine into competitive advantage.

Where Pith is reading between the lines

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

  • Management practices that already support human knowledge sharing may transfer directly to AI integration without wholesale redesign.
  • Firms could test whether AI tools producing more intuitive outputs integrate faster into the spiral than purely analytical ones.
  • The model implies that economic returns to AI adoption depend on organizational context rather than on the technology alone.

Load-bearing premise

That AI outputs function as tacit machine knowledge directly analogous to human tacit knowledge so that the existing conversion processes and need for shared organizational context apply unchanged.

What would settle it

An empirical case where AI-generated knowledge fails to convert into organizational knowledge through the standard socialization-externalization-combination-internalization steps without requiring entirely new validation or integration mechanisms.

read the original abstract

Nonaka emphasized that innovation is the result of a continuous back-and-forth between tacit and explicit knowledge. Artificial intelligence introduces a fundamentally new object into this process -- tacit machine knowledge -- but Nonaka's ideas are more relevant than ever. The central role of the knowledge-creating company remains the same: to create the shared context in which different kinds of knowledge can feed off each other, become organizational knowledge, and set off further cycles of innovation.

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 / 1 minor

Summary. The manuscript claims that artificial intelligence introduces a new category of knowledge—'tacit machine knowledge'—into Nonaka's SECI spiral of knowledge conversion between tacit and explicit forms, yet the central role of the knowledge-creating company in establishing shared contexts for organizational knowledge and innovation cycles remains unchanged.

Significance. If the core analogy holds, the work would suggest that existing organizational knowledge-management frameworks can absorb AI outputs without structural revision, offering a conceptual bridge between Nonaka's model and contemporary AI deployment in firms. The absence of empirical tests, formal derivations, or falsifiable predictions, however, confines its contribution to reframing rather than advancing testable theory.

major comments (2)
  1. [Abstract] Abstract: The central assertion that AI outputs constitute 'tacit machine knowledge' directly analogous to human tacit knowledge (and therefore participate in the four SECI conversion modes) is introduced without a definition of the properties that render machine outputs 'tacit' (embodied, context-dependent, difficult to articulate) or any mechanism showing entry into or exit from those modes.
  2. [Abstract] Abstract and concluding discussion: The claim that 'the central role of the knowledge-creating company remains the same' rests on the unargued premise that no new validation, representation, or conversion constructs are required; the text supplies neither an illustration of how machine outputs traverse the spiral nor a contrast case demonstrating when the shared-context requirement would fail.
minor comments (1)
  1. The term 'tacit machine knowledge' is used without prior clarification of its relation to existing distinctions between explicit AI artifacts and human tacit knowledge, which reduces precision when the argument is extended to organizational practice.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We respond to each major comment below and indicate where revisions will be made to clarify the conceptual extension.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central assertion that AI outputs constitute 'tacit machine knowledge' directly analogous to human tacit knowledge (and therefore participate in the four SECI conversion modes) is introduced without a definition of the properties that render machine outputs 'tacit' (embodied, context-dependent, difficult to articulate) or any mechanism showing entry into or exit from those modes.

    Authors: We agree the abstract would benefit from an explicit definition. The manuscript characterizes tacit machine knowledge as outputs embodied in model parameters (analogous to human embodiment), dependent on training data and prompt context, and difficult to articulate without further processing or human interpretation. We will revise the abstract to include this characterization and add a short paragraph illustrating entry via socialization (human-AI interaction) and exit via externalization (AI output documented as explicit knowledge). revision: yes

  2. Referee: [Abstract] Abstract and concluding discussion: The claim that 'the central role of the knowledge-creating company remains the same' rests on the unargued premise that no new validation, representation, or conversion constructs are required; the text supplies neither an illustration of how machine outputs traverse the spiral nor a contrast case demonstrating when the shared-context requirement would fail.

    Authors: The argument is that machine outputs still require human sensemaking and organizational embedding to become shared knowledge, so the company's role in providing shared context is unchanged. We will add a concrete illustration of traversal (e.g., AI-generated prototypes socialized in cross-functional teams) and a contrast case where absence of shared context causes AI outputs to remain unintegrated and fail to trigger further innovation cycles. revision: yes

Circularity Check

1 steps flagged

Unexamined analogy treating AI outputs as 'tacit machine knowledge' directly interchangeable with human tacit knowledge in the SECI model

specific steps
  1. self definitional [Abstract]
    "Artificial intelligence introduces a fundamentally new object into this process -- tacit machine knowledge -- but Nonaka's ideas are more relevant than ever. The central role of the knowledge-creating company remains the same: to create the shared context in which different kinds of knowledge can feed off each other, become organizational knowledge, and set off further cycles of innovation."

    The new object is introduced by name as 'tacit machine knowledge' and immediately declared to preserve the original spiral structure and organizational role. The conclusion that no new mechanisms are required is therefore entailed by the labeling itself rather than by any demonstrated equivalence in embodiment, articulation difficulty, or conversion dynamics.

full rationale

The paper's core claim is that AI adds 'tacit machine knowledge' yet leaves Nonaka's SECI spiral and the knowledge-creating company's role unchanged. This follows directly from labeling AI outputs with the same 'tacit' descriptor used for human knowledge, without supplying an independent definition of machine tacitness, a mechanism for its participation in the four conversion modes, or external validation that the shared-context requirement transfers. The provided abstract text exhibits the reduction; no further equations or self-citations appear in the excerpt.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 1 invented entities

The paper introduces one new postulated entity without independent evidence or derivation. No free parameters or standard axioms are invoked beyond the background Nonaka framework.

invented entities (1)
  • tacit machine knowledge no independent evidence
    purpose: To extend Nonaka's tacit-explicit conversion model to include AI as a knowledge source
    Presented as a fundamentally new object in the knowledge spiral; no falsifiable handle or external evidence is supplied in the abstract.

pith-pipeline@v0.9.1-grok · 5591 in / 1226 out tokens · 42984 ms · 2026-06-30T02:40:31.654672+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

49 extracted references · 17 canonical work pages

  1. [1]

    Prometheus , year =

    Sanzogni, Louis and Guzman, Gustavo and Busch, Peter , title =. Prometheus , year =

  2. [2]

    Organization Studies , year =

    Ribeiro, Rodrigo and Collins, Harry , title =. Organization Studies , year =

  3. [3]

    and Mousavi, Reza , title =

    Alavi, Maryam and Leidner, Dorothy E. and Mousavi, Reza , title =. Journal of the Association for Information Systems , year =. doi:10.17705/1jais.00859 , url =

    work page doi:10.17705/1jais.00859

  4. [4]

    Kybernetes , year =

    Sumbal, Muhammad Saleem and Amber, Quratulain , title =. Kybernetes , year =. doi:10.1108/K-06-2023-1126 , url =

    work page doi:10.1108/k-06-2023-1126 2023

  5. [5]

    VINE Journal of Information and Knowledge Management Systems , year =

    Böhm, Karsten and Durst, Susanne , title =. VINE Journal of Information and Knowledge Management Systems , year =. doi:10.1108/VJIKMS-10-2024-0357 , url =

    work page doi:10.1108/vjikms-10-2024-0357 2024

  6. [6]

    Managing Human and Artificial Knowledge , editor =

    Kirchner, Kathrin and Scarso, Enrico , title =. Managing Human and Artificial Knowledge , editor =. 2026 , pages =. doi:10.1007/978-3-032-14721-9_6 , url =

    work page doi:10.1007/978-3-032-14721-9_6 2026

  7. [7]

    2026 , note=

    Uchihira, Naoshi , title =. 2026 , note=. doi:10.48550/arXiv.2603.21866 , url =. 2603.21866 , archivePrefix =

    work page doi:10.48550/arxiv.2603.21866 2026

  8. [8]

    Agents and Multi-agent Systems: Technologies and Applications 2024 , editor =

    Matsumoto, Takashi and Nishikawa, Ryu and Morimoto, Chikako , title =. Agents and Multi-agent Systems: Technologies and Applications 2024 , editor =. 2025 , pages =. doi:10.1007/978-981-97-6469-3_12 , url =

    work page doi:10.1007/978-981-97-6469-3_12 2024

  9. [9]

    Journal of Knowledge Management , year =

    He, Xiaomei and Antonczak, Laurent and Burger-Helmchen, Thierry , title =. Journal of Knowledge Management , year =. doi:10.1108/JKM-07-2025-1065 , url =

    work page doi:10.1108/jkm-07-2025-1065 2025

  10. [10]

    Journal of Innovation & Knowledge , year =

    Cerchione, Roberto and Liccardo, Giuseppe and Passaro, Renato , title =. Journal of Innovation & Knowledge , year =. doi:10.1016/j.jik.2025.100866 , url =

    work page doi:10.1016/j.jik.2025.100866 2025

  11. [11]

    Organization Science , year =

    Epple, Dennis and Argote, Linda and Devadas, Rukmini , title =. Organization Science , year =. doi:10.1287/orsc.2.1.58 , url =

    work page doi:10.1287/orsc.2.1.58

  12. [12]

    Organization Science , year =

    Kogut, Bruce and Zander, Udo , title =. Organization Science , year =. doi:10.1287/orsc.3.3.383 , url =

    work page doi:10.1287/orsc.3.3.383

  13. [13]

    Organization Science , year =

    Zander, Udo and Kogut, Bruce , title =. Organization Science , year =. doi:10.1287/orsc.6.1.76 , url =

    work page doi:10.1287/orsc.6.1.76

  14. [14]

    Organization Science , year =

    Kogut, Bruce and Zander, Udo , title =. Organization Science , year =. doi:10.1287/orsc.7.5.502 , url =

    work page doi:10.1287/orsc.7.5.502

  15. [15]

    , title =

    Grant, Robert M. , title =. Strategic Management Journal , year =. doi:10.1002/smj.4250171110 , url =

    work page doi:10.1002/smj.4250171110

  16. [16]

    Argote, Linda , title =

  17. [17]

    Nonaka, Ikujiro and Takeuchi, Hirotaka , title =

  18. [18]

    California Management Review , year =

    Nonaka, Ikujiro and Konno, Noboru , title =. California Management Review , year =. doi:10.2307/41165942 , url =

    work page doi:10.2307/41165942

  19. [19]

    Organization Science , year =

    Nonaka, Ikujiro and von Krogh, Georg , title =. Organization Science , year =. doi:10.1287/orsc.1080.0412 , url =

    work page doi:10.1287/orsc.1080.0412

  20. [20]

    Organization Science , year =

    Brown, John Seely and Duguid, Paul , title =. Organization Science , year =. doi:10.1287/orsc.2.1.40 , url =

    work page doi:10.1287/orsc.2.1.40

  21. [21]

    2025 , institution=

    How People Use ChatGPT , author=. 2025 , institution=

    2025

  22. [22]

    1991 , doi =

    Lave, Jean and Wenger, Etienne , title =. 1991 , doi =

    1991

  23. [23]

    Ide, Enrique and Talamàs, Eduard , title =

  24. [24]

    Beraja, Martin and Talamàs, Eduard , title =

  25. [25]

    The Quarterly Journal of Economics , volume=

    Generative AI at work , author=. The Quarterly Journal of Economics , volume=. 2025 , publisher=

    2025

  26. [26]

    2026 , month =

    Johnston, Drew and Holtz, David and Richmond, Alex Martin and Ong, Christopher and Tambe, Prasanna and Chatterji, Aaron , title =. 2026 , month =

    2026

  27. [27]

    Ouyang, Long and Wu, Jeff and Jiang, Xu and Almeida, Diogo and Wainwright, Carroll L. and Mishkin, Pamela and Zhang, Chong and Agarwal, Sandhini and Slama, Katarina and Ray, Alex and Schulman, John and Hilton, Jacob and Kelton, Fraser and Miller, Luke and Simens, Maddie and Askell, Amanda and Welinder, Peter and Christiano, Paul and Leike, Jan and Lowe, R...

    2022

  28. [28]

    2022 , month = nov, url =

    Introducing. 2022 , month = nov, url =

    2022

  29. [29]

    Mathematical methods and human thought in the age of AI.arXiv preprint arXiv:2603.26524, March

    Klowden, Tanya and Tao, Terence , title =. 2026 , note =. 2603.26524 , archiveprefix=

    work page arXiv 2026

  30. [30]

    Journal of Political Economy , year =

    Ide, Enrique and Talamàs, Eduard , title =. Journal of Political Economy , year =

  31. [31]

    The Economics of Transformative

    Chatterji, Aaron and Rock, Daniel and Talamàs, Eduard , title =. The Economics of Transformative. 2025 , chapter =

    2025

  32. [32]

    Polanyi, Michael , title =

  33. [33]

    Harvard Business Review , year =

    Nonaka, Ikujiro , title =. Harvard Business Review , year =

  34. [34]

    Organization Science , year =

    Nonaka, Ikujiro , title =. Organization Science , year =

  35. [35]

    Science , year =

    Brynjolfsson, Erik and Mitchell, Tom , title =. Science , year =

  36. [36]

    AEA Papers and Proceedings , year =

    Brynjolfsson, Erik and Mitchell, Tom and Rock, Daniel , title =. AEA Papers and Proceedings , year =

  37. [37]

    , title =

    Autor, David H. , title =. 2024 , doi =

    2024

  38. [38]

    The UNESCO Courier , year =

    Juez, Beatriz , title =. The UNESCO Courier , year =

  39. [39]

    McCarthy, Dan , title =

  40. [40]

    Let's Take a Look at How

  41. [41]

    Systems and Methods to Mimic Target Food Items Using Artificial Intelligence , howpublished =

    Pichara, Karim and Zamora, Pablo and Muchnick, Mat. Systems and Methods to Mimic Target Food Items Using Artificial Intelligence , howpublished =

  42. [42]

    Danley, Sam , title =

  43. [43]

    Hall, Christine , title =

  44. [44]

    Global Food-Tech Unicorn

  45. [45]

    Silver, David and Huang, Aja and Maddison, Chris J. and Guez, Arthur and Sifre, Laurent and van den Driessche, George and Schrittwieser, Julian and Antonoglou, Ioannis and Panneershelvam, Veda and Lanctot, Marc and Dieleman, Sander and Grewe, Dominik and Nham, John and Kalchbrenner, Nal and Sutskever, Ilya and Lillicrap, Timothy and Leach, Madeleine and K...

  46. [46]

    , title =

    Shin, Minkyu and Kim, Jin and van Opheusden, Bas and Griffiths, Thomas L. , title =. Proceedings of the National Academy of Sciences , year =

  47. [47]

    and Yaghi, Omar M

    Zheng, Zhiling and Rong, Zichao and Rampal, Nakul and Borgs, Christian and Chayes, Jennifer T. and Yaghi, Omar M. , title =. Angewandte Chemie International Edition , year =

  48. [48]

    Science , year =

    Mosalaganti, Shyamal and Obarska-Kosinska, Agnieszka and Siggel, Marc and Taniguchi, Reiya and Turo. Science , year =

  49. [49]

    Barry Callebaut Partners with