0 cumulative citations
View corpus contextWorkers are quietly self-teaching with generative AI in a process the paper terms 'shadow learning', shifting knowledge and power out of managerial control; firms that ignore these informal practices risk underestimating productivity gains and misdesigning training and governance policies.
Shadow Learning and Knowledge Redistribution: How Workers Appropriate Generative <scp>AI</scp> Outside Formal Organisational Systems
openalex
theoretical
n/a evidence
8/10 relevance
Summary only
summary available; pdf_status=paywall
DOI
Source
PDF
Shadow learning is the informal, worker-led appropriation of generative AI that reshapes organisational learning processes and redistributes knowledge authority, and should be treated as a constitutive feature rather than an aberration of formal training systems.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
0 cumulative citations
View corpus contextABSTRACT Organisations are ceding visibility over how their employees develop knowledge. As generative AI tools become widely accessible, workers are steadily building capabilities and co‐producing knowledge through informal AI engagements that operate entirely outside sanctioned training systems. This paper theorises shadow learning, the informal and autonomous appropriation of AI by workers for knowledge expansion beyond managerial design, as a process that fundamentally reconfigures how knowledge authority is distributed in organisations. Integrating critical management learning with sociomaterial and post‐human perspectives, and drawing on current debates in AI ethics and human–AI interaction, the paper conceptualises shadow learning through a recursive model of constraint, appropriation and reconfiguration. This framework shows how AI functions as a sociomaterial catalyst that alters who holds knowledge authority and how learning unfolds in practice. The paper makes three contributions. It introduces shadow learning as a concept that extends beyond informal learning by foregrounding its subversive and sociomaterial dimensions; it bridges critical and sociomaterial perspectives to show how AI‐mediated learning works at once as political resistance and distributed practice and it draws both traditions into contact with current debates in AI ethics. The contribution lies in that integration and its application to a phenomenon neither tradition has theorised alone, not in restating a post‐human case sociomaterial scholarship has already made. The model offers knowledge management scholars and practitioners a conceptual tool for understanding why informal AI‐mediated knowledge practices are not deviations from organisational learning systems but constitutive features of them.
Summary
Main Finding
The paper introduces and theorises "shadow learning": the informal, autonomous appropriation of generative AI by workers to build knowledge outside sanctioned training systems. Shadow learning is framed as a sociomaterial, often subversive process that redistributes knowledge authority within organisations. Rather than being deviations from formal learning, AI-mediated informal practices are constitutive features of organisational learning systems.
Key Points
- Definition: Shadow learning — informal, worker-led learning using AI tools that operates outside managerial design and control.
- Sociomaterial framing: AI is treated as an actor or catalyst that co-constitutes learning practices, not merely a neutral tool.
- Recursive model: Learning dynamics are theorised through recursive stages of constraint (formal systems & managerial control), appropriation (workers adopting AI for their own learning), and reconfiguration (changes in knowledge authority and practice).
- Political dimension: Shadow learning can function as a form of resistance to managerial knowledge control, shifting power over expertise.
- Conceptual contribution: Bridges critical management learning and sociomaterial/post-human perspectives, and connects them to AI ethics and human–AI interaction debates.
- Practical claim: Informal AI-mediated learning should be understood as integral to organisational learning, with implications for knowledge management and governance.
- Limitations: The paper is conceptual/theoretical — it develops a model and argument but does not present new empirical data.
Data & Methods
- Approach: Theoretical synthesis and conceptual modelling.
- Sources integrated: Critical management learning literature; sociomaterial and post-human theory; debates in AI ethics; human–AI interaction research.
- Method: Construction of a recursive conceptual model (constraint → appropriation → reconfiguration) to explain how generative AI reshapes learning and knowledge authority in organisations.
- Empirical status: No original quantitative or qualitative data reported; the contribution is analytical and integrative rather than empirical.
Implications for AI Economics
- Human capital formation: Shadow learning accelerates informal human capital accumulation, which may change the returns to firm-provided training and the value of on-the-job experience.
- Productivity measurement: Productivity gains from AI adoption may be underestimated if informal, distributed learning outside formal channels is not observed or credited to firms.
- Incentives and investment: Firms may underinvest in formal training when workers self-upskill via accessible AI tools, altering optimal firm training strategies and contracting over skill development.
- Knowledge externalities and appropriation: Shadow learning creates positive externalities (diffusion of skills) and potential tensions over IP, tacit knowledge capture, and compensating workers for firm-relevant skill acquisition.
- Labor bargaining and wage dynamics: Redistribution of knowledge authority can strengthen workers’ bargaining positions if skills become more portable, potentially affecting wage premia for AI-augmented skills.
- Organizational boundaries and firm strategy: Firms might need to rethink governance, monitoring, and incentive systems (rather than trying to suppress informal AI use) and design architectures that integrate or channel shadow learning constructively.
- Regulation and policy: Policy debates (privacy, IP, safety, workplace surveillance) intersect with shadow learning — regulators should consider how restrictions on tool access or monitoring could suppress productive informal learning or shift its distribution.
- Research agenda for AI economics:
- Empirically measure prevalence of shadow learning and its contribution to productivity.
- Estimate returns to AI-mediated informal learning versus formal training.
- Study how firm policies (tool provisioning, monitoring, incentives) shape the scale and effects of shadow learning.
- Analyze distributional consequences across occupations, skill levels, and firm sizes.
Overall, the paper suggests that AI economics should incorporate informal, distributed learning mediated by generative AI into models of human capital, firm behavior, and productivity, and that policies or firm strategies ignoring shadow learning risk mischaracterising the diffusion and impacts of AI.
Assessment
Paper Typetheoretical
Evidence Strengthn/a — The paper is conceptual and synthesizes existing literatures; it presents a theoretical model without original empirical data, so there is no empirical causal evidence to evaluate.
Methods Rigormedium — The paper offers a careful, interdisciplinary synthesis and a clear recursive model (constraint → appropriation → reconfiguration), drawing on critical management, sociomaterial theory, AI ethics, and HCI; however, it lacks empirical testing, robustness checks, or formal modelling that would strengthen causal claims.
SampleNo original sample or dataset; the paper is based on theoretical synthesis of prior literature in critical management learning, sociomaterial/post-human theory, AI ethics, and human–AI interaction research.
Themesskills_training human_ai_collab org_design productivity adoption
GeneralizabilityNo empirical evidence — prevalence and effects of shadow learning are unmeasured and may vary widely by sector, occupation, and country., Dynamics likely differ across firm size, regulatory environments, tool access, and digital literacy; theoretical claims may not hold uniformly., Conceptual framing may emphasize certain organizational contexts (e.g., knowledge work) and be less applicable to manual or tightly regulated work., Cultural and institutional differences (labor institutions, training norms) could alter the political dynamics described.
Claims (9)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| The paper defines “shadow learning” as informal, worker-led learning with generative AI that occurs outside managerial design and control. Training Effectiveness | positive | Informal AI-mediated learning outside formal organisational training systems |
Reading fidelity
high
Study strength
low
|
not reported
|
| The paper argues that generative AI is not merely a neutral learning tool but acts as an actor or catalyst that co-constitutes organisational learning practices. Training Effectiveness | positive | The organisation and formation of learning practices |
Reading fidelity
high
Study strength
low
|
not reported
|
| The paper theorises shadow learning through a recursive sequence of constraint, appropriation, and reconfiguration. Organizational Efficiency | null_result | Changes in learning practices and knowledge authority |
Reading fidelity
high
Study strength
low
|
not reported
|
| The paper argues that shadow learning can operate as resistance to managerial control over knowledge and can redistribute authority over expertise within organisations. Governance And Regulation | positive | Distribution of knowledge authority and control over expertise |
Reading fidelity
high
Study strength
low
|
not reported
|
| The paper argues that AI-mediated informal learning should be treated as a constitutive part of organisational learning systems rather than as a deviation from formal learning. Training Effectiveness | positive | Recognition and integration of informal learning in organisational learning systems |
Reading fidelity
high
Study strength
low
|
not reported
|
| The paper proposes that shadow learning may accelerate informal human-capital accumulation and thereby change the returns to firm-provided training and on-the-job experience. Skill Acquisition | positive | Informal skill acquisition and returns to formal training and experience |
Reading fidelity
high
Study strength
speculative
|
not reported
|
| The paper suggests that measured productivity gains from AI adoption may be underestimated when informal, distributed learning outside formal channels is not observed or attributed to firms. Firm Productivity | positive | Measurement of productivity gains associated with AI adoption |
Reading fidelity
high
Study strength
speculative
|
not reported
|
| The paper argues that firms should reconsider governance, monitoring, and incentive systems and seek to integrate or channel shadow learning rather than simply suppress informal AI use. Governance And Regulation | positive | Organisational governance of informal AI-mediated learning |
Reading fidelity
high
Study strength
speculative
|
not reported
|
| The paper is conceptual and theoretical and does not report original quantitative or qualitative empirical data. Other | null_result | Empirical evidence status of the paper |
Reading fidelity
high
Study strength
high
|
not reported
|