The shadow digital economy is a structured, marketized criminal ecosystem whose modular labour markets and reusable tooling let illicit activity scale—and AI is lowering costs and quality barriers. Regulators should treat AI‑enabled illicit markets as a systemic policy concern and fold them into financial and consumer‑protection frameworks.

Main Finding

The paper conceptualizes the shadow digital economy (SDE) as a distinct, expanding sector where illegal, informal, and concealed digital activities (from malware development to transnational data trading and platform‑based labor exploitation) form integrated value chains. It maps the SDE’s labor‑market structure, identifies institutional traps that sustain shadow employment, proposes a taxonomy of criminal digital products/services and a “shadow project” model, and argues the SDE imposes broad economic harms beyond direct monetary loss (erosion of trust, reputational damage, national‑security risks). The authors call for periodic assessment and explicit integration of AI‑related risks into financial and governance frameworks.

Key Points

• Definition and scope
  • SDE: economic activity using digital technologies that violates law or regulation, is concealed from authorities, and forms a distinct system of relations and value chains.
  • Uses OECD categories (underground, illegal, informal, household production, statistical underground) as a base for classification.
• Labor‑market roles and structure
  • Identified participant types: researchers (vulnerability discovery), developers (malware/tools), vendors/distributors, IT technicians/hosters, hackers/exploiters, fraudsters/social engineers, managers/coordinators.
  • Labor is often dual (formal employment + shadow contracting); tasks can be fragmented so many workers lack awareness of illegality.
  • Platform hiring and remote contractor models facilitate cross‑border, hard‑to‑trace labor.
• Institutional traps that sustain SDE
  • Legalization trap: normalization/stabilization of illegal practices.
  • Trust trap: internal trust/arbitration mechanisms make networks resilient.
  • Corruption trap: dependencies on corrupt channels hinder reform.
  • Automatism trap: routine/legacy practices persist despite risks.
• Economic and social impacts
  • Beyond direct financial losses: undermines consumer trust, damages brand reputation via breaches/fraud/counterfeiting, erodes labor quality, reduces tax/social contributions, and poses regional/national security threats.
  • Shadow employment causes underpayment, unsafe working conditions, and reduces incentives for productivity and technology adoption.
• Conceptual and practical contributions
  • Presents a concise “shadow project” model detailing roles/composition of criminal projects.
  • Proposes classification of criminally oriented digital products/services and highlights specialized services meriting further research.
• Data limitations
  • No unified methodology or robust statistical data exists; underreporting by firms and governments hinders measurement.

Data & Methods

• Approach: qualitative/systematic literature review and structural synthesis.
• Sources: scholarly and practitioner literature, security‑industry reports; databases searched included Google Scholar, IEEE Xplore, and Scopus for 2019–2024 (plus a few other publications).
• Frameworks used: OECD typology for non‑observed economy; thematic categorization into technological risks, digital economy definitions, labor markets, shadow/non‑observed/criminal economies.
• Outputs: conceptual definitions, labor‑market typology, institutional trap analysis, taxonomy of criminal products/services, and the “shadow project” model.
• Limitations: lack of primary empirical data, reliance on secondary sources and reports, acknowledged underreporting and measurement challenges for SDE activity.

Implications for AI Economics

• AI as a dual accelerant: generative and automation AI can both enable SDE (automating phishing, malware generation, synthetic identity creation, scaling fraud) and strengthen defenses (automated detection, threat intelligence). Economic models must account for these opposing forces.
• Labor‑market effects
  • Skill composition and measurement: AI may shift demand toward high‑skill cyber roles while enabling decentralized, platform‑based contracting that obscures employment status; standard labor statistics may understate activity and earnings linked to SDE.
  • Dual employment and gigification: AI tools lower coordination costs for shadow projects, increasing prevalence of multi‑affiliation workers and complicating employer/employee definitions used in economic analysis.
• Measurement and macroeconomic accounting
  • GDP and productivity estimates risk bias if SDE‑related economic activity grows but remains unmeasured; AI‑enabled illicit commerce could amplify this distortion.
  • Financial stability: AI‑driven cyber incidents can produce systemic shocks (payment disruptions, large‑scale fraud), so macroprudential frameworks should include cyber/AI risk scenarios.
• Policy and regulatory implications
  • Integrate AI‑related cybercrime risks into financial governance, AML/CTF, and tax policy; require periodic evaluations including AI risk vectors.
  • Improve incident reporting (mandatory breach and loss disclosure) and cross‑border information sharing to reduce underreporting.
  • Regulate and monitor the misuse of AI tools (e.g., models that enable automated scam generation) while fostering responsible AI standards in the security sector.
  • Invest in workforce policies: upskilling/reskilling, cybersecurity education, incentives to reduce shadow employment.
• Research and monitoring priorities
  • Develop indicators and methodologies to estimate SDE size and AI’s role within it.
  • Study AI‑enabled specialized SDE services (automated fraud-as-a-service, synthetic persona marketplaces, AI‑generated deepfakes used for extortion).
  • Model feedback loops between institutional traps and AI adoption in both legitimate firms and SDE actors.
• Recommended actions for economists and policymakers
  • Incorporate SDE and AI cyber‑risk externalities into models of productivity, labor supply, and fiscal revenues.
  • Adopt mandatory, standardized incident and breach reporting across jurisdictions; strengthen international cooperation for policing cross‑border AI‑enabled illicit markets.
  • Design targeted interventions to break institutional traps (anti‑corruption measures, digital sovereignty investments, rehabilitation pathways for marginal participants).

Short takeaway: the SDE is a structurally organized, digitally enabled shadow sector that reshapes labor markets and imposes broad economic costs. AI multiplies both risks and mitigation options; AI economics should explicitly model and monitor AI‑driven illicit activity, adjust measurement frameworks, and guide policy to integrate AI‑related cyber risk into financial and labor governance.