ADOPTION OF ARTIFICIAL INTELLIGENCE IN THE RUSSIAN EXTRACTIVE INDUSTRY: ANALYSIS OF BARRIERS AND INSTITUTIONAL CONSTRAINTS

This article presents an analysis of the adoption of artificial intelligence technologies in the extractive industry of the Russian Federation, with a focus on identifying barriers and institutional constraints that hinder their diffusion and scaling. The study is based on a comparison of digitalization levels and AI investment volumes in Russia and leading global economies — the United States, China, and the European Union — over the period 2020–2025. The findings show that Russia’s lag in AI adoption in the extractive sector is systemic in nature and is associated with limited access to data, insufficient investment volumes, as well as human capital and infrastructure constraints. The study concludes that a comprehensive approach is required to develop an institutional environment conducive to the adoption of artificial intelligence in the extractive industry of the Russian Federation

Summary

Main Finding

Russia’s lag in adopting AI within its extractive industries is systemic. Between 2020–2025 Russia trails the United States, China and the EU on both digitalization indicators and AI investment volumes for the mining/oil & gas sectors. The gap reflects not only smaller investment flows but also institutional constraints — limited data access, weak data governance, human capital shortages, and inadequate digital infrastructure — which together suppress diffusion and scaling of AI applications. Addressing the problem requires a coordinated, institutional-level response rather than isolated technical fixes.

Key Points

Comparative trend (2020–2025): the US, China and EU show materially higher digitalization levels and AI investment per unit of extractive output than Russia; Russia’s adoption is both slower and shallower.
Data access is a primary bottleneck: fragmented, proprietary or closed datasets, unclear ownership rules, and weak mechanisms for safe data sharing hinder model training and cross-firm applications.
Investment insufficiency: absolute and relative AI investment volumes in the Russian extractive sector are lower; private risk capital is limited and public support is not sufficiently targeted to scale-up projects.
Human capital constraints: shortages of AI talent in industry-specific roles, limited retraining of engineering staff, and brain drain reduce absorption capacity.
Infrastructure shortfalls: sensorization, connectivity (edge/cloud), computing hardware and localized software stacks are underdeveloped relative to peers.
Institutional failures: weak standards/interoperability, limited public–private coordination, regulatory uncertainty, and sanctions/import restrictions that affect access to hardware/software exacerbate diffusion problems.
The problem is systemic: the barriers interact (e.g., lack of data reduces demand for talent; weak infrastructure deters investment), so piecemeal measures will have limited effect.

Data & Methods

Comparative multi-country analysis over 2020–2025 using publicly available investment and digitalization indicators for extractive industries (national/industry statistics, investment datasets and sectoral digitalization indices).
Metrics compared include AI investment volumes (public + private, where reported), digitalization/digit maturity proxies in extractive sectors, and supporting indicators (ICT infrastructure, workforce/education statistics, patent/activity counts and project case evidence).
Institutional analysis: review of data governance frameworks, regulatory regimes, standards and market structure across jurisdictions, with emphasis on how these shape data access, procurement and public–private collaboration.
Augmentation by qualitative evidence: firm-level case studies and expert commentary (where available) to illustrate concrete adoption barriers and scaling failures.
Analytical approach: cross-country trend comparison, identification of co-moving constraints, and synthesis of institutional factors likely to explain observed investment/adoption gaps. (The study focuses on structural and institutional interpretation rather than causal identification from randomized interventions.)

Implications for AI Economics

Institutional determinants matter: diffusion models of AI should explicitly incorporate institutional variables (data governance, standards, public infrastructure) alongside traditional human-capital and capital-cost channels. These institutional constraints generate persistent adoption frictions and can explain sustained international gaps.
Market failures likely justify public action: data externalities, coordination failures, and large fixed costs of sensorization/computing imply underinvestment from private actors alone. Well-designed public interventions (data platforms, co-financing, standards) can have high social returns.
Policy interventions should be comprehensive and sequenced: combine (a) unlocking data (clear ownership, safe-sharing frameworks, interoperable formats), (b) targeted investment incentives (matching grants, procurement commitments, risk-sharing for pilots), (c) human-capital programs (upskilling, industry–university links), and (d) core infrastructure (sensors, connectivity, local compute capacity). Fragmented policies are unlikely to overcome interacting bottlenecks.
Strategic and geopolitical considerations: sanctions and supply-chain limits affect hardware & software availability; domestic substitution or international cooperation channels will alter adoption paths and costs. Economic models should capture how geopolitical risk changes the price and availability of AI capital.
Research directions: quantify welfare gains from specific AI applications in extraction (productivity, safety, emissions), evaluate cost-effectiveness of different policy bundles, and estimate dynamic returns to building data ecosystems and human capital that enable broader GPT-like spillovers across sectors.

Assessment

Paper Typedescriptive Evidence Strengthmedium — Uses multiple publicly available quantitative indicators (investment volumes, digitalization proxies, infrastructure and workforce stats) across countries and complements them with firm case studies and institutional review, which gives convergent descriptive evidence of a gap; however, it lacks causal identification, may suffer from measurement gaps and reporting heterogeneity across jurisdictions, and does not estimate causal effects of AI on economic outcomes. Methods Rigormedium — Combines cross-country trend comparison, diverse indicator sources and qualitative case evidence and discusses institutional mechanisms, but relies on aggregate and proxy metrics with potential comparability issues, limited transparency on data construction/aggregation, and no quasi-experimental or counterfactual strategy to isolate causal drivers. SampleComparative multi-country dataset spanning 2020–2025 covering extractive industries (mining, oil & gas) in Russia, the United States, China and the EU using national and industry statistics, investment datasets (public + private where reported), sectoral digitalization indices, ICT infrastructure measures, workforce/education statistics, patent/activity counts, supplemented by firm-level case studies and expert commentary. Themesadoption governance productivity human_ai_collab GeneralizabilityFocused on extractive industries (mining, oil & gas); findings may not apply to services or manufacturing sectors., Comparisons are limited to Russia vs US/China/EU; results may not generalize to other emerging economies or subnational regions., Relies on public and reported investment data and proxies which may differ in coverage and definition across countries (measurement/comparability issues)., Time window 2020–2025; dynamics may change rapidly with policy shifts or geopolitical events (e.g., sanctions, supply shocks)., Qualitative case studies are illustrative and not necessarily representative of all firms within each jurisdiction.

Claims (14)

Claim	Direction	Confidence	Outcome	Details
Between 2020–2025 Russia trails the United States, China and the EU on both digitalization indicators and AI investment volumes in the mining and oil & gas sectors. Adoption Rate	negative	medium	digitalization levels and AI investment volumes per unit of extractive output (mining and oil & gas)	0.11
Russia’s adoption of AI in extractive industries is both slower (lower growth rate) and shallower (lower depth of digitalization) than peer jurisdictions in 2020–2025. Adoption Rate	negative	medium	rate of change in digitalization indicators and depth of digitalization (digit maturity proxies) in extractive industries	0.11
The gap is driven not only by smaller investment flows but also by institutional constraints—limited data access, weak data governance, human capital shortages, and inadequate digital infrastructure—that together suppress diffusion and scaling of AI applications. Adoption Rate	negative	medium	diffusion and scaling of AI applications in extractive industries	0.11
Data access is a primary bottleneck: datasets are fragmented, often proprietary or closed, ownership rules are unclear, and mechanisms for safe data sharing are weak, hindering model training and cross-firm applications. Governance And Regulation	negative	medium	availability and usability of industrial data for AI model training and cross-firm applications	0.11
Absolute and relative AI investment volumes in the Russian extractive sector are lower than in the US, China and EU; private risk capital is limited and public support insufficiently targeted to scale-up projects. Adoption Rate	negative	medium	AI investment volumes (absolute and per unit of extractive output); availability of private risk capital and targeted public support	0.11
There are human capital constraints: shortages of AI talent in industry-specific roles, limited retraining of engineering staff, and brain drain reduce the sector's capacity to absorb and deploy AI. Skill Acquisition	negative	medium	industry-specific AI talent supply, retraining rates for engineering staff, measured absorption capacity for AI projects	0.11
Infrastructure shortfalls — insufficient sensorization, limited connectivity (edge/cloud), inadequate computing hardware and immature localized software stacks — are underdeveloped in Russia relative to peers and hinder deployment. Adoption Rate	negative	medium	sensor density, connectivity quality (edge/cloud readiness), availability of computing hardware and local software stacks for AI deployment	0.11
Institutional failures—weak standards/interoperability, limited public–private coordination, regulatory uncertainty, and sanctions/import restrictions—exacerbate diffusion problems for AI in extractive sectors. Governance And Regulation	negative	medium	standards/interoperability quality, level of public–private coordination, regulatory clarity, access to hardware/software	0.11
The barriers to AI adoption in Russia’s extractive industries interact systemically (e.g., lack of data reduces demand for talent; weak infrastructure deters investment), so piecemeal measures will have limited effect. Governance And Regulation	negative	medium	overall effectiveness of isolated vs. coordinated interventions on AI diffusion and scaling	0.11
Institutional determinants (data governance, standards, public infrastructure) materially influence AI diffusion and should be incorporated explicitly into diffusion models alongside human capital and capital-cost channels. Research Productivity	positive	medium	model explanatory power for AI diffusion when including institutional variables vs. excluding them	0.11
Market failures—data externalities, coordination failures, and large fixed costs for sensorization/computing—likely lead to underinvestment by private actors and justify targeted public interventions (data platforms, co-financing, standards). Market Structure	positive	medium	degree of private underinvestment in AI enabling assets and projected social returns from public interventions	0.11
Sanctions and supply-chain restrictions affect access to hardware and software, altering adoption paths and increasing costs; domestic substitution or international cooperation will influence future trajectories. Adoption Rate	negative	medium	availability and cost of hardware/software inputs for AI and resulting adoption trajectories	0.11
Effective policy should be comprehensive and sequenced: unlock data (clear ownership, safe-sharing frameworks), provide targeted investment incentives (matching grants, procurement commitments), run human-capital programs (upskilling, industry–university links), and build core infrastructure (sensors, connectivity, local compute). Governance And Regulation	positive	speculative	improvement in AI diffusion, scaling, and impact in extractive sectors resulting from coordinated policy bundles	0.02
Research gaps remain: quantifying welfare gains from specific AI applications in extraction (productivity, safety, emissions), evaluating cost-effectiveness of policy bundles, and estimating dynamic returns to data ecosystems and human capital. Research Productivity	null_result	high	magnitude of welfare gains from AI applications; cost-effectiveness metrics for policy interventions; dynamic returns to data and human capital investments	0.18

Entities

Russia (population) Extractive industries (mining; oil and gas) (population) AI investment volumes (public and private) (outcome) Digitalization / digitization maturity proxies (outcome) Digitalization indicators (outcome) Data access (outcome) Data governance frameworks (institution) Human capital shortages (AI talent / engineering staff) (population) Infrastructure shortfalls (sensorization, connectivity, compute, software stacks) (outcome) United States (institution) China (institution) European Union (institution) Sensorization (outcome) Connectivity (edge and cloud) (outcome) Computing hardware (outcome) Comparative multi-country analysis (2020–2025) (method) Institutional analysis (method) Cross-country trend comparison (method) Public–private coordination (institution) Regulatory uncertainty and regimes (institution) Standards and interoperability (institution) Sanctions and import restrictions (institution) Private risk capital / private sector finance (institution) Public sector support / public financing (institution) Firm-level case studies (method) National and industry statistics (dataset) Investment datasets (dataset) Sectoral digitalization indices (dataset) Patent and activity counts (outcome) GPT-like models (ai_tool) Productivity; Safety; Emissions (outcome) Market failures (data externalities, coordination failures, fixed costs) (outcome) Data platforms (policy instrument) (institution) Matching grants (policy instrument) (institution) Upskilling programs and industry–university links (institution)