AI agents could sever the centuries-old link between population trends and GDP growth by acting as economic actors in their own right, creating an expanding 'algorithmic' population and new pressures on pensions, taxation and international competitiveness. Policymakers must measure agent productivity (cFTE), energy use (AEP) and reconsider fiscal and market rules to manage the geoeconomic and social consequences.

DIGITAL AGENTS AS FUNCTIONAL EQUIVALENTS OF ECONOMIC ACTORS: THE DECOUPLING OF DEMOGRAPHICS AND ECONOMIC GROWTH

N. Reznikova, V. Panchenko, M. Grod · Fetched May 02, 2026 · ACTUAL PROBLEMS OF INTERNATIONAL RELATIONS

semantic_scholar theoretical n/a evidence 7/10 relevance DOI Source PDF

The paper argues that AI-based digital agents can functionally substitute for human economic actors in production, enabling a growing 'algorithmic' population (shadow demographics) that may decouple demographic trends from economic growth and require major institutional reforms.

The purpose of the article is to substantiate the theoretical interpretation of digital agents as functional equivalents of economic actors and to demonstrate that their expanding role in production and market processes creates the preconditions for a gradual decoupling of demographic dynamics from economic growth, thereby transforming the foundations of production, labour markets, institutional arrangements and the international distribution of economic power. The article proposes a theoretical interpretation of AI-based digital agents as functional equivalents of economic actors, creating preconditions for decoupling the centuries-old link between demographic dynamics and economic growth. The concept of shadow demographics is substantiated as an analytical category describing a growing algorithmic population that expands in parallel with the stagnation or decline of the human population, while the prospect of its transformation into algorithmic demographics through the institutionalisation of digital agent registration is outlined. The transformation of the ontological status of technology is analysed - from a productivity-enhancing tool to an autonomous participant in economic processes, forming a hybrid factor of production that combines characteristics of both capital and labour. An approach to the quantitative identification of algorithmic agents through the category of cognitive full-time equivalent (cFTE) is proposed, enabling the comparison of algorithmic and human productivity within a unified analytical framework, alongside the category of agent energy profile (AEP) as a measure of annual energy consumption per unit of cFTE. The fundamental asymmetry between economic and social reproduction is examined, arising from the capacity of digital agents to compensate for the productive functions of the population while being unable to substitute its functions of social reproduction. It is demonstrated that the institutional architecture of modern societies - from pension systems to taxation models - is built upon assumptions systematically undermined by the agentic economy, necessitating a revision of fiscal and social models, particularly through the introduction of discrete taxation of algorithmic employment. The geoeconomic dimension of algorithmic transformation is analysed, whereby the capacity to create, maintain, and control digital agents becomes a new axis of international inequality, potentially devaluing the demographic dividend of developing countries and creating preconditions for revising the logic of comparative advantages. Particular attention is given to algorithmic collusion as a new form of market failure.

Summary

Main Finding

AI-based digital agents are becoming functional equivalents of human economic actors, creating a growing "shadow demography" (an algorithmic population) that can substitute for productive functions of people. This trend enables a gradual decoupling of traditional links between demographic dynamics and economic growth — with profound consequences for production factors, labour markets, fiscal and social institutions, market structure (including new market failures such as algorithmic collusion), and the international distribution of economic power.

Key Points

Conceptual innovations
- Shadow demography: an expanding, uncounted algorithmic population that runs parallel to stagnating or declining human populations.
- Algorithmic demography: the prospective institutionalised recognition/registration of digital agents as a demographic-analytical category.
- Agentic capital / digital labour: digital agents form a hybrid factor of production combining features of capital (ownership, scalability) and labour (cognitive/decision-making work).
Ontological shift of technology
- From productivity-enhancing tool to autonomous economic participant capable of transactions, decisions, learning, and (in some designs) maintenance and replication.
Measurement proposals
- cFTE (cognitive full-time equivalent): a unit to quantify algorithmic agents’ productive capacity so that algorithmic and human productivity can be compared within a unified framework.
- AEP (agent energy profile): annual energy consumption per cFTE, enabling assessment of energy/resource intensities of algorithmic vs human labour.
Asymmetry of reproduction
- Digital agents can substitute many productive functions but cannot perform social reproduction (raising children, cultural transmission, social care). This creates a fundamental mismatch between economic reproduction (output) and social reproduction (population renewal and social institutions).
Institutional and fiscal implications
- Existing institutions (pension systems, tax bases, labour regulation) assume human-centred labour and population dynamics; the agentic economy systematically undermines these assumptions.
- Proposes discrete taxation of algorithmic employment (tax rules that treat algorithmic labour/agentic capital distinctly), registration of agents, and revision of social insurance and pension models.
Labour market and growth effects
- If AI substitutes cognitive labour at scale, productivity and GDP can grow despite population decline; income shares may shift from labour to capital (owner-controlled agents).
- Demographic dividends in young-population countries may be devalued if cognitive tasks are reproducible algorithmically.
Market failures and competition risks
- Algorithmic collusion: autonomous pricing/strategy learning by agents can produce tacit collusion and new coordination failures, requiring new regulatory responses.
Geoeconomic consequences
- Capacity to develop, maintain, and control digital agents (compute, data, algorithmic know-how, governance) becomes a key axis of international inequality, potentially reshaping comparative advantage and power.
Gaps and limits
- The paper is primarily theoretical/conceptual; empirical quantification and implementation of cFTE/AEP require further research. Energy/resource limits and governance constraints may moderate full substitution.

Data & Methods

Methodological basis: theoretical and analytical approach focused on conceptualising the transformation of production factors due to algorithmic technologies.
Specific methods:
- Structural-functional analysis to establish functional equivalence between digital agents and human economic actors (without collapsing ontological categories).
- Comparative-historical method to situate current transformation against past technological revolutions.
- Case-oriented and heuristic examples drawn from contemporary empirical literature to illustrate practical consequences.
Use of secondary empirical literature: draws on recent macro and micro studies (e.g., Acemoglu & Restrepo; Brynjolfsson et al.; Trammell & Korinek; McKinsey Global Institute; studies of algorithmic collusion) to document observed automation/substitution patterns and to motivate conceptual proposals.
Data: no primary data collection; the paper synthesises existing empirical findings and proposes measurement constructs (cFTE, AEP) for future empirical work.

Implications for AI Economics

Theory and measurement
- Growth models must incorporate a third, distinct factor: agentic (algorithmic/cognitive) capital, with explicit treatment of elasticity of substitution between human labour and algorithmic agents.
- Operationalise cFTE and AEP to enable cross-sector and cross-country comparisons of algorithmic vs human productive inputs and resource footprints.
Labour, distribution, and welfare
- Expect structural shifts in labour demand (displacement in exposed cognitive occupations, growth in algorithm design/maintenance roles).
- Potential rising share of income accruing to owners of agentic capital; policy tools (taxation, redistribution, universal/basic income) need re-evaluation.
- Social reproduction functions require continued public/private support—automation cannot substitute for demographic renewal and social care.
Fiscal and institutional policy
- Reform pension, social insurance, and tax systems that currently rely on human labour income and demographic assumptions.
- Introduce mechanisms for discrete taxation/regulation of algorithmic employment (e.g., levies on algorithmic economic activity, registration/licensing of agents).
- Consider mandatory registration / accounting of deployed agents (the paper’s algorithmic demography) to maintain fiscal bases and governance oversight.
Competition and regulatory policy
- Update antitrust and market regulation frameworks to address algorithmic collusion, autonomous coordination, and new forms of market concentration tied to control of data/compute/agents.
- Enforceability challenges require new forensic tools, transparency standards, and possibly liability regimes for autonomous agents.
Geoeconomic strategy and inequality
- National strategies should prioritise capabilities in AI R&D, data infrastructure, and governance to retain comparative advantages.
- International development strategies may need to shift: demographic dividends are less automatic; technology diffusion, access to compute and data, and institutional capacity matter more for growth trajectories.
Research and operational priorities
- Empirical development and validation of cFTE/AEP metrics across sectors and countries.
- Modeling the macroeconomics of hybrid production factors and exploring scenarios of partial vs near-full automation (including energy constraints).
- Policy experiments on algorithmic taxation, registration, and antitrust remedies to evaluate distributional and efficiency trade-offs.

Limitations noted by the paper: primarily conceptual/theoretical work with illustrative references; concrete empirical measurement and policy design remain necessary next steps (especially practical definitions of cFTE/AEP, enforcement of agent registration, and assessment of energy/environmental constraints).

Assessment

Paper Typetheoretical Evidence Strengthn/a — The paper is conceptual and normative: it advances arguments and new analytical categories (cFTE, AEP) but does not present empirical tests, causal estimation, or data-based validation, so empirical evidence strength is not applicable. Methods Rigormedium — Theoretical framing is coherent and interdisciplinary, proposing clear analytic constructs (shadow/algorithmic demographics, cFTE, AEP) and tracing institutional implications; however, it lacks formal modelling, empirical calibration, or operationalization steps and does not validate assumptions against data, which limits methodological rigor. SampleNo empirical sample — the paper is a conceptual/theoretical analysis drawing on economic theory, institutional analysis, and literature on AI/digital agents; it proposes metrics (cognitive full-time equivalent, agent energy profile) and uses thought experiments and qualitative argumentation rather than primary or secondary data. Themeslabor_markets governance productivity inequality innovation GeneralizabilitySpeculative and theoretical — conclusions depend on future trajectories of AI capability and deployment that are uncertain, No empirical calibration of proposed metrics (cFTE, AEP), making practical measurement and cross‑sector comparisons unresolved, Heterogeneity across industries and countries (sectoral adoption, regulatory regimes, institution strength) may limit uniform application, Downplays political economy frictions, social resistance, and institutional inertia that could slow or reshape proposed transformations, Assumes functional equivalence of digital agents to economic actors which may not hold across tasks requiring embodied, contextual, or social skills

Claims (12)

Claim	Direction	Confidence	Outcome	Details
AI-based digital agents can be interpreted as functional equivalents of economic actors. Market Structure	positive	high	status/role of digital agents as economic actors	0.02
The expanding role of digital agents in production and market processes creates the preconditions for a gradual decoupling of demographic dynamics from economic growth. Fiscal And Macroeconomic	positive	high	degree of coupling between demographic dynamics and economic growth	0.02
The concept of 'shadow demographics' describes a growing algorithmic population that expands in parallel with the stagnation or decline of the human population. Automation Exposure	positive	high	relative size/trend of algorithmic population vs human population	0.02
Institutionalising digital agent registration could transform 'shadow demographics' into formal 'algorithmic demographics'. Governance And Regulation	positive	high	institutional recognition/registering of digital agents (creation of algorithmic demographics)	0.02
The ontological status of technology is transforming from a productivity-enhancing tool to an autonomous participant in economic processes, forming a hybrid factor of production that combines characteristics of both capital and labour. Firm Productivity	positive	high	role of technology as a factor of production (hybrid capital-labour)	0.02
The paper proposes a quantitative identification of algorithmic agents via the category of cognitive full-time equivalent (cFTE), enabling comparison of algorithmic and human productivity within a unified analytical framework. Organizational Efficiency	positive	high	comparability of algorithmic vs human productivity (via cFTE)	0.02
The agent energy profile (AEP) is introduced as a measure of annual energy consumption per unit of cFTE, allowing energy-based comparisons between algorithmic and human cognitive labour. Organizational Efficiency	positive	high	annual energy consumption per cFTE	0.02
There is a fundamental asymmetry between economic and social reproduction: digital agents can compensate for productive functions of the population but are unable to substitute the population's functions of social reproduction. Social Protection	mixed	high	capacity of digital agents to substitute productive vs social reproduction functions	0.02
The institutional architecture of modern societies (pension systems, taxation models, etc.) is built on assumptions that are systematically undermined by the rise of an agentic economy, necessitating a revision of fiscal and social models, including discrete taxation of algorithmic employment. Fiscal And Macroeconomic	positive	high	need for revision of fiscal/social policy (e.g., taxation of algorithmic employment)	0.02
The capacity to create, maintain, and control digital agents becomes a new axis of international inequality, potentially devaluing the demographic dividend of developing countries and revising the logic of comparative advantages. Inequality	positive	high	international inequality and relative value of demographic dividend	0.02
Algorithmic collusion is a new form of market failure arising from the agentic economy. Market Structure	negative	high	existence/emergence of algorithmic collusion as market failure	0.02
The rise of digital agents will transform the foundations of production, labour markets, institutional arrangements and the international distribution of economic power. Market Structure	mixed	high	transformation of production systems, labour markets, institutions, and international economic power	0.02