Extracting candidates for peer-reviewed papers: a cross-wiki analysis that excludes the author's own theory — Notes

A research-strategy note that extracted 20 candidates at the grain of a peer-reviewed paper, examining four clusters (epistemology and methodology, education and learning, profession and industry, evaluation and creativity) and limited to the themes among the 117 wiki/note entries that do not depend on the author's own…

Purpose and method

This note extracts, from the research notes accumulated in wiki/note, the themes among the author’s interests that are at the grain of a peer-reviewed paper. Candidates that presuppose, extend, or test the theory that the author is building separately were excluded. We divided the 117 notes into four clusters (13 on epistemology and methodology, 15 on education and learning, 20 on profession and industry, and 20 on evaluation and creativity), and an independent subagent read every note in each cluster in full. In the judgment, we first separated plain literature summaries from the author’s own claims, syntheses, and methods, and turned only the original ones into candidates.

Cross-cutting finding

Even with the author’s own theory set aside, the interests of the notes gather around one center of gravity. That center is the science of evaluation and exploration. It has four faces: (1) what can be measured and how (LLM-as-a-Judge, the context dependence of capability, competency measurement), (2) the conditions under which measurement and substitution break learning (cognitive offloading, failure-driven learning), (3) the sociology of professions that are hard to measure (emerging occupations, stagnant pay, the gap between adoption and evaluation), and (4) how novelty in research itself is made (metascience). Whichever candidate is chosen, this center of gravity allows it to be presented as part of a consistent research identity of the author.

Tier 1: nearly ready to write (maturity A / A−)

#	Working title	Type	Basis notes	Target venues
2	What offloading breaks learning? (a framework for judging alignment with learning goals)	Integrative review + theory	Patterns of Technological Harm to Education, and Their Refutations: A Robustness Assessment Across 25 Works, Does Early Use of Generative AI Inhibit the Formation of Thought? A Literature Map of Cognitive Offloading and Learning, Designing Failure Into Learning: From Expectation Failure to Productive Failure and Industry Implementations	Educational Psychology Review, Educational Research Review
3	Non-classical context dependence of AI capability measurement (a test design for a relational ontology)	Position paper	Is Capability Inside the Model? A Relational Ontology of AI Evaluation and a Test for Contextuality, AI Research Gaps, Revisited: Sorting Symmetry-Completion from Coverage, and Digging Out an Endogenous Scaling Law	Philosophy & Technology, Minds and Machines, ICML/NeurIPS position
4	Strategies for establishing novelty: a two-layer classification of generating and claiming	Integrative review	How Novelty Is Made and How It Is Claimed: A Typology of Establishing Strategies, Where Do Novel Research Questions Come From? The Four Loci and Their Combinations, Examined Against the Literature	Research Evaluation, Quantitative Science Studies
5	The academia-industry gap in Agentic Experience	Scoping review	AX and Design: A Cross-Comparison of Academic and Industry Perspectives, The Academic Foundations of Agentic Experience (AX): Where Human-Agent Interaction Research Stands, Industry Discourse on Agentic Experience (AX): What Vendors, Consultants, and Practitioners Are Saying	CHI, ToCHI, IJHCS

For Candidate 2, a triple corpus of about 70 sources (original harm claims, peer-reviewed refutations, and the debate on failure design) and a robustness judgment table are in place. The main contribution is a synthesis that passes a single criterion (“harm occurs when the cognitive process being substituted is itself the current learning goal”) through 50 years of harm discourse. Candidate 3 is the only candidate that has passed several acceptance devices (the operating sheets A to G, rewriting of the board, and methods from outside the family). It is feasible because the test apparatus for sheaf-contextuality has already been applied to word senses in LLMs in Proc. R. Soc. A. For Candidate 4, a corpus of 243 items and a ledger of 135 items exist, and the main work is the formalization of the review methodology. Candidate 5 is nearly complete in content, but the field becomes outdated quickly, so keeping it current would be the limiting factor if it were submitted.

Tier 2: skeleton complete, reinforcement needed (maturity B)

#	Working title	Type	Basis notes	Main gaps
6	A jurisdictional analysis of the forward deployed designer	Document analysis + concept	Can the Forward Deployed Designer Stand as a Role?, Design and the Professions: A Literature Map of Design's Position Seen Through the Theory of Professions	Expanding the corpus of job postings, interviews with 5 to 10 people
7	A six-factor model of stagnant design pay	Integrative review	Structural Factors behind Designer Pay Stagnation — Academic Review (2026), Structural Factors behind Designer Pay Stagnation — Industry Trends (2026), The Economics of AI and Design — Reading the Structural Transformation of Design Labor Through Six Economic Theories	Refining the causal order, stating the methodology
8	”Using but not endorsing”: a structural model of the gap between AI adoption and evaluation	Theory paper	90% Adoption x 10% Approval — The Paradox of AI Tool Diffusion and Evaluative Divergence, From Maker to Editor: A Structural Analysis of the Designer Role Transition in the Age of AI	Formalizing propositions, checking sources at first hand
9	Generative AI strikes directly at the learning goals of design education	Position paper	Issues for Design Education (Not Craft Education): From the Evidence on Failure Design and Cognitive Offloading, Does Early Use of Generative AI Inhibit the Formation of Thought? A Literature Map of Cognitive Offloading and Learning	Checking the canon of pedagogy in the original, adding prior work on studio pedagogy

#	Working title	Type	Basis notes	Main gaps
11	Can evaluation frameworks of computational creativity evaluate generative AI?	Integrative review	The Scholarly Lineage of Generative Art: A 94-Item Literature Map from Information Aesthetics to the Post-LLM Era, Which Angle for a Generative-Art Review Paper Attracts Citations? An Empirical Analysis Using Bibliographic Data, Adversarial Review of Two Generative Art Surveys: Examining the Literature Map and the Citation Strategy Against the Design-Theory Canon	Additional collection on the side of evaluation frameworks, 18 unverified bibliographic records
12	Is subtraction a generative operation symmetric with addition? (an audit of additive bias in creativity measures)	Theory + desk audit	The Creativity of Subtraction: The Additive Bias, and the Hypothesis of Designing Absence	Systematic check of the claim that no prior work exists, carrying out the audit
13	Do combinations of novelty dimensions raise evaluations? (a direct test of combinations of seats)	Bibliometrics	Where Do Novel Research Questions Come From? The Four Loci and Their Combinations, Examined Against the Literature	All of corpus construction and analysis
14	The undiscovery rate: a quantitative indicator of knowledge loss	Theory + pilot measurement	The Vanishing of Knowledge: Undiscovery as the Dual of Discovery	Mathematical formulation of the indicator, measurement on public data
15	Is co-evolution driven by cognition or by evaluative infrastructure? (an empirical test of temporal precedence)	Empirical (theory testing)	Connecting design theory and sociology: where does novelty get lost?, デザイン研究がまだ使っていない社会学、すでに使い尽くした社会学private	Field access, research design, ethics review
16	A scoping review of seven types of purpose in the educational use of generative art	Scoping review	What Is Generative Art For in Education? Seven Purpose Types and a Literature Map of 72 Studies	Restructuring to comply with PRISMA-ScR

Tier 3: idea stage, with implementation or experiments still entirely ahead (maturity C)

#	Working title	Type	Basis notes	Main gaps
17	Endogenous scaling laws (a closed-loop formulation of the ratio of synthetic data)	Theory + computational experiment	AI Research Gaps, Revisited: Sorting Symmetry-Completion from Coverage, and Digging Out an Endogenous Scaling Law, AI Research Gaps, a Third Time: Rereading Them as Rewritings of the Game Board	Mathematical model and experiments over repeated generations
18	Is the design system the context layer of AI agents?	Concept + empirical	Design Systems as AI's Foundation Layer — Structured Design Knowledge Determines Agent Accuracy, The Genealogy of Generative AI Engineering and Its Assessment from a Design Perspective	Operational definition of maturity and a direct test experiment
19	A system design education tool that fully implements Goal-Based Scenarios with LLMs	System paper	Scenario-Based System Design Learning: An Industry Survey in the Context of Goal-Based Scenarios, Prior Research on Goal-Based Scenarios: A 31-Item Literature Map from Schank's Theory to Components, Evidence, and Design Contexts	Tool implementation, evaluation design, participants
20	A comparative experiment that measures the sign of the usefulness of hallucination	Empirical (experiment)	An Academic Map of Research That Treats Hallucination as a Resource for Creativity, Recombination of the Known and Nonhuman Reasoning: Can Hallucination Be Treated as a Resource for Creativity?	Experimental design, a human baseline, models of several generations

Recommended order of submission

For the second paper, Candidate 2 (the framework for judging alignment with learning goals) is suitable. It is the only education-related candidate rated A that can move into writing without additional experiments, and it also gives the theoretical basis for Candidates 9, 10, and 19, so it supports the papers that follow. If one more paper is prepared in parallel, Candidate 3 (the relational ontology) is the high-risk, high-return choice. As a position paper complete in theory alone, it can enter the submission line, and if it succeeds, it can intervene in the assumptions of AI evaluation research itself. If the decision is based on time sensitivity, Candidate 6 (the forward deployed designer) loses value the longer it waits, because it holds data that cannot be reproduced and that captures how an occupational title came into being (including job postings that have since disappeared).

Narrowing to qualitative research and field-testing novelty (added 2026-08-09)

Of the 20 candidates, 5 can stand as qualitative research (interviews, document analysis, participant observation, qualitative content analysis). For these 5, we did not rely on the notes’ own statements that nothing was found. Instead, we conducted a novelty audit that identified the closest prior work through external literature searches. The judgment has three levels: intact, survives with limits, and substantially pre-empted.

Rank	Candidate	Qualitative method	Novelty judgment	Limiting factor
2	A jurisdictional analysis of the forward deployed designer (Candidate 6)	Document analysis + interviews	Survives with limits	Interviews with 5 to 10 people
3	Temporal precedence of the drivers of co-evolution (Candidate 15)	Process study with trace data + participant observation	Survives with limits	Field access
4	The interaction between generative AI and the crit (recast from Candidate 9)	Participant observation and discourse analysis of crits	Survives only after reframing	Access to studios
5	Managing practices of “using but not endorsing” (recast from Candidate 8)	Semi-structured interviews	Substantially pre-empted (recasting required)	Rebuilding the design for differentiation

Candidate 6: the forward deployed designer

No academic study even of the FDE, let alone the FDD, was identified, so the literal combination stands. However, every component has appeared separately: the professionalization discourse of UX in Kou and Gray 2018, job postings × UX × generative AI in Honorato and colleagues 2026, job-posting analysis of emerging AI occupational titles in research on prompt engineers, and professional autonomy and boundaries × AI × qualitative research in medical sociology (radiologists). To defend it, the study should target the success-case bias of the sociology of professions, which studies occupations only after their markets have formed, and should be designed to judge the claims of an occupational title in real time, before institutionalization. The fact that both job postings disappeared is not a weakness; it becomes observational data on a jurisdictional claim that can fail. A comparative case design with the FDE (which reached market recognition) would make a theoretical contribution beyond the description of a single case.

Candidate 15: the drivers of co-evolution

The combination of three points (treating durable evaluative infrastructure as a candidate driver, and identifying its precedence relative to cognitive updating with time-stamped trace data) was not found, and the core of the design remains. However, the claims of absence were too broad in two places. For “co-evolution models assume a process within cognition,” the social extensions by Wiltschnig and colleagues 2013 and by Parsons and colleagues are counterexamples, and “no study has mapped Espeland and Sauder onto design” is refuted by the DIS ‘24 paper on reactivity (which applies Espeland and Sauder’s reactivity to metrics that track warehouse workers’ labor and discusses design directions). The largest threat is Harrison and Rouse 2015 (which has shown empirically that external evaluative interactions co-construct the problem space). The study incorporates it in the text as the bearer of a rival hypothesis and is recast as a three-way adjudication that discriminates three accounts (co-evolution within cognition, interpersonal interaction, and durable evaluative infrastructure) with data on time differences.

Candidate 9 recast: the interaction between generative AI and the crit

The original design of interviewing educators about their perceptions is already occupied by Fleischmann’s studies (students’ experiences in 2024 and interviews with nine educators in 2026) and several studies in architectural education, and carrying it out as planned would be reviewed as a replication. The gap that remains is a description of the interaction in actual review sessions into which AI-generated work has been brought: how educators probe, confirm, and reconstruct traces of the students’ own exploration. Novelty stands if the unit of analysis moves from perception to the interaction of the crit and the theoretical lens is placed in the lineage of failure design (the integration phase of productive failure).

Candidate 8 recast: “using but not endorsing”

Müller and colleagues 2026 have already carried out a qualitative study of the gap among knowledge workers, and pairs of components (designers × professional identity, creatives × the autonomy paradox) were also filled separately between 2024 and 2026. “The first integration of three theories” is a weak claim of the gap-spotting type and cannot be defended in peer review. Three paths remain. (1) Ask, in explicit contrast with Müller and colleagues, whether designers’ non-endorsement is rooted in stakes specific to creative work, namely the attribution of authorship and aesthetic judgment (a boundary condition of the mechanism); (2) move the question from the existence of the gap to the practices of managing it (disguising use, selective disclosure); (3) draw value for comparative cases from the rarity of the context of Japanese design organizations.

Overall ranking as qualitative research

Candidate 6 ranks second, with time-limited data that cannot be reproduced and a clear path to reframing. Candidate 15 ranks third; its theoretical return is the largest, but field access is the limiting factor. Candidates 9 and 8 stand only after recasting, and they come later in the order of work. For Candidates 6, 8, and 9, carrying out the study in the Japanese context is an independent source of differentiation from existing research, which is concentrated in the United States, Europe, Australia, and China.

Follow-up work

In response to the conclusion of this audit that gap-spotting novelty claims cannot be defended, assumption-reversing question generation was carried out in Problematizing the wiki corpus: an assumption ledger for four domains and twelve challenges (12 challenges developed from 45 assumptions in four domains). Writing Material for a Review Article: The Assumption Ledger and Nearest-Neighbour Literature Left by 17 Novelty Audits rearranges the audit results of both notes as material for writing a review paper. On 2026-10-11, a scan in A Scan for Research Gaps in Past Notes (2026-10, Fifteen Lenses) divided the ways of finding gaps into 15 lenses, applied them to the research rows of the corpus, searched for the closest existing research for each candidate, and judged it (10 candidates from 87 notes).

Unverified items

The novelty audit created the following seven items that require checking.

The methodological details of the jury observation in AI & SOCIETY 2024
The method and scope of Fleischmann 2026
The publication status of the Research Square preprint on desirable difficulties
The AMJ paper on workplace jurisdiction (full text not read)
Confirming the theoretical framework of Honorato and colleagues 2026 in the published version
The authors and analytic details of the 2024 Design Studies paper on power asymmetries
How far designers are included in the qualitative part of Müller and colleagues 2026

Each requires a first-hand check before the corresponding candidate is submitted. The audit also relied on general web searches, and systematic searches of Scopus, Web of Science, and the ACM DL and a sweep of literature in languages other than English have not been carried out.

Unverified markers inherited from the basis notes also remain. They are 1 check of a venue’s acceptance record for Candidate 1, 3 bibliographic records for Candidate 3, more than ten bibliographic records for Candidate 4, 11 primary sources for Candidate 6, the majority of citations for Candidate 8, 6 works of the pedagogy canon for Candidate 9, 18 bibliographic records for Candidate 11, 4 original sources for Candidate 12, 2 bibliographic records for Candidate 14, and 3 bibliographic records for Candidate 17. This analysis itself created 0 new unverified items.

References

This list gives the external literature named in the text, with bibliographic records checked against Crossref, arXiv, and publishers’ pages. The basis notes are linked from the tables in the text and are not listed here.

Candidate 3: the test apparatus for contextuality

Lo, K. I., Sadrzadeh, M. & Mansfield, S. (2025). Quantum-like contextuality in large language models. Proceedings of the Royal Society A 481(2319), 20240399. https://doi.org/10.1098/rspa.2024.0399

Candidate 6: emerging occupational titles and professional jurisdiction

Kou, Y. & Gray, C. M. (2018). Towards professionalization in an online community of emerging occupation: Discourses among UX practitioners. GROUP ‘18, 322–334. https://doi.org/10.1145/3148330.3148352
Honorato, F., Sosa, R. & Clemente, V. (2026). UX design meets GenAI: Emerging practices from job postings in Singapore. Base Diseño e Innovación 11(13). https://doi.org/10.52611/bdi.num13.2026.1753
Vu, A. & Oppenlaender, J. (2025). Prompt engineer: Analyzing hard and soft skill requirements in the AI job market. arXiv:2506.00058. https://arxiv.org/abs/2506.00058
Lombi, L. & Rossero, E. (2024). How artificial intelligence is reshaping the autonomy and boundary work of radiologists: A qualitative study. Sociology of Health & Illness 46(2), 200–218. https://doi.org/10.1111/1467-9566.13702
Kim, B., Rezazade Mehrizi, M. H. & Bailey, D. E. (2025). Interlacing situated and algorithmic modes of knowledge work: A workplace jurisdiction perspective. Academy of Management Journal 68(5), 907–938. https://doi.org/10.5465/amj.2023.1286

Candidate 15: the drivers of co-evolution

Wiltschnig, S., Christensen, B. T. & Ball, L. J. (2013). Collaborative problem–solution co-evolution in creative design. Design Studies 34(5), 515–542. https://doi.org/10.1016/j.destud.2013.01.002
Parsons, P. C. & Ridley, A. (2025). Judgment as coordination: A joint systems view of visualization design practice. arXiv:2507.01209. https://arxiv.org/abs/2507.01209
Parsons, P. C. & Shukla, P. C. (2025). Beyond problem solving: Framing and problem-solution co-evolution in data visualization design. arXiv:2508.07058. https://arxiv.org/abs/2508.07058
Zhang, W. & Ma, J. (2024). Framing under power asymmetries: A cross-level examination of the early-stage product design process. Design Studies 91–92, 101250. https://doi.org/10.1016/j.destud.2024.101250
Cheon, E. (2024). Examining algorithmic metrics and their effects through the lens of reactivity. DIS ‘24, 3179–3192. https://doi.org/10.1145/3643834.3660676
Harrison, S. H. & Rouse, E. D. (2015). An inductive study of feedback interactions over the course of creative projects. Academy of Management Journal 58(2), 375–404. https://doi.org/10.5465/amj.2012.0737

Candidate 9 recast: design education and generative AI

Fleischmann, K. (2024). The commodification of creativity: Integrating generative artificial intelligence in higher education design curriculum. Innovations in Education and Teaching International 62(6), 1843–1857. https://doi.org/10.1080/14703297.2024.2427039
Fleischmann, K. (2026). From tools to thinking partners: Cognitive and pedagogical shifts in design education through generative AI. Arts and Humanities in Higher Education 25(4), 340–361. https://doi.org/10.1177/14740222261420495
Karadağ, D. (2025). AI in architectural education: Rethinking studio culture. PLANARCH 9(2), 243 ff. https://doi.org/10.54864/planarch.1749891
Iranmanesh, A. & Lotfabadi, P. (2024). Critical questions on the emergence of text-to-image artificial intelligence in architectural design pedagogy. AI & Society 40(5), 3557–3571. https://doi.org/10.1007/s00146-024-02111-x

Candidate 8 recast: the gap between AI adoption and attitudes

Müller, M., dos Santos, N. E. G., Kubatova, J. & Mesquita Cardoso, L. (2026). Navigating cognitive dissonance in generative AI adoption: Ethical and psychological dimensions of knowledge worker acceptance. European Journal of Innovation Management. https://doi.org/10.1108/EJIM-03-2025-0305
Li, J., Cao, H., Lin, L., Hou, Y., Zhu, R. et al. (2024). User experience design professionals’ perceptions of generative artificial intelligence. CHI ‘24. https://doi.org/10.1145/3613904.3642114
Zulu, N. C., Åkesson, M. & Thomsen, M. (2026). What does generative AI mean for my professional competence? Insights into the perspectives of digital creative professionals. Interacting with Computers. https://doi.org/10.1093/iwc/iwaf056
Tang, C. (2026). The autonomy paradox in AI-generated content adoption: Creative-specific alternative to TAM model in China’s micro-short drama industry. PLOS One 21(1), e0336166. https://doi.org/10.1371/journal.pone.0336166
Vasconcelos, D. (2026). Resisting AI: A conceptual framework for practices of resistance against artificial intelligence in the creative industries. AI & Society 41(9), 8445–8460. https://doi.org/10.1007/s00146-026-03135-1
Fang, C., Guo, J., Khiatani, P. V., Lin, H., Zhang, B. et al. (2026). Beyond one-size-fits-all: Profiling designers’ adoption of generative artificial intelligence in professional design practice. International Journal of Human–Computer Interaction. https://doi.org/10.1080/10447318.2026.2658255