Evidence (3308 claims)
Search and filter individual claims pulled from the papers. Looking for a specific finding ("what's the effect on wages?"), you're in the right place. Want to compare whole outcome categories against each other instead? Use the Evidence Explorer.
The board below groups claims two ways: by broad theme (nine paper-level topics) and by outcome category (the 34 claim-level outcomes that the Explorer and Syntheses also use).
Browse by theme
Nine broad, paper-level topics. Click one to filter the claims below.
Adoption
9875 claims
Filter claims →
Productivity
8807 claims
Filter claims →
Governance
7870 claims
Filter claims →
Human-AI Collaboration
7560 claims
Filter claims →
Org Design
4892 claims
Filter claims →
Innovation
4781 claims
Filter claims →
Labor Markets
4004 claims
Filter claims →
Skills & Training
3308 claims
Filtered →
Inequality
2332 claims
Filter claims →
Claims by outcome category
Counts by direction of finding. These are the same 34 outcome categories the Explorer compares and the Syntheses are written for. A linked row has a published synthesis.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 870 | 233 | 116 | 1066 | 2363 |
| Governance & Regulation | 976 | 451 | 218 | 133 | 1809 |
| Organizational Efficiency | 949 | 224 | 144 | 88 | 1416 |
| Technology Adoption Rate | 764 | 287 | 141 | 122 | 1325 |
| Research Productivity | 501 | 152 | 74 | 362 | 1101 |
| Output Quality | 542 | 216 | 69 | 69 | 896 |
| Decision Quality | 387 | 198 | 94 | 54 | 740 |
| Firm Productivity | 513 | 67 | 101 | 27 | 714 |
| AI Safety & Ethics | 249 | 303 | 73 | 36 | 667 |
| Market Structure | 190 | 192 | 134 | 27 | 548 |
| Task Allocation | 243 | 77 | 91 | 36 | 452 |
| Innovation Output | 291 | 33 | 55 | 20 | 401 |
| Skill Acquisition | 206 | 72 | 65 | 21 | 364 |
| Employment Level | 133 | 63 | 115 | 22 | 335 |
| Fiscal & Macroeconomic | 153 | 79 | 52 | 32 | 323 |
| Task Completion Time | 206 | 37 | 12 | 15 | 272 |
| Firm Revenue | 179 | 52 | 29 | 5 | 266 |
| Consumer Welfare | 130 | 76 | 47 | 13 | 266 |
| Inequality Measures | 48 | 137 | 51 | 6 | 242 |
| Worker Satisfaction | 101 | 81 | 25 | 13 | 220 |
| Error Rate | 84 | 110 | 11 | 5 | 210 |
| Wages & Compensation | 98 | 47 | 30 | 10 | 185 |
| Regulatory Compliance | 88 | 73 | 17 | 7 | 185 |
| Automation Exposure | 66 | 64 | 33 | 16 | 182 |
| Team Performance | 105 | 29 | 30 | 11 | 176 |
| Training Effectiveness | 109 | 22 | 14 | 21 | 168 |
| Developer Productivity | 114 | 21 | 14 | 8 | 158 |
| Job Displacement | 12 | 90 | 24 | 1 | 127 |
| Hiring & Recruitment | 57 | 9 | 9 | 5 | 80 |
| Skill Obsolescence | 6 | 56 | 9 | 1 | 72 |
| Social Protection | 43 | 17 | 8 | 2 | 70 |
| Creative Output | 35 | 21 | 9 | 4 | 70 |
| Labor Share of Income | 18 | 21 | 17 | 1 | 57 |
| Worker Turnover | 15 | 16 | — | 4 | 35 |
| Industry | — | — | — | 1 | 1 |
Skills Training
Remove filter
Even after expanded university output plus non-degree routes, a persistent shortage remains that will signal upward pressure on wages for in-demand AI skills.
Combined coverage measured at 43.9% of estimated demand and observed wage differentials in the monitoring data; authors infer labor-supply constraint and wage pressure from shortfall and wage observations.
On the metric of training volume, universities have broadly complied with the Russian Government’s directive to expand AI specialist training.
Reported increases/levels of AI-related program enrollments and graduate numbers across the 191 monitored institutions compared to the government directive target (paper’s policy conclusion based on program volume data).
A practical policy framework for an inclusive transition should: diagnose exposure, protect affected workers, prepare the workforce (education and lifelong learning), promote human-augmenting adoption, and monitor & iterate using data and evaluations.
Policy synthesis based on comparative institutional analysis, empirical program evaluations where available, and theoretical guidance on complementarities and reallocation.
Policy interventions—investment in lifelong learning, active labor market policies, social protection, and incentives for equitable AI deployment—can reduce adverse distributional impacts and make the transition more inclusive.
Synthesis of theoretical frameworks and empirical evaluations of targeted programs (training, wage subsidies, portable benefits) where quasi-experimental or experimental evidence exists; comparative policy analysis.
Alternative social-insurance architectures (partial prefunding, universal transfers, UBI-style schemes financed by K_T rents) can mitigate social strains arising from declining payroll bases, according to simulated scenarios.
Calibrated model policy simulations exploring prefunded pensions, universal transfers, and financing mechanisms using captured rents from K_T; comparisons of pension sustainability and welfare outcomes across scenarios.
Shifting part of the tax burden from labor to returns on K_T (corporate, property, rent, or wealth taxes) can help restore revenue bases and internalize displacement externalities, but such measures face avoidance, evasion, and international coordination challenges.
Policy experiments in the structural model showing effects of capital/wealth taxation on fiscal balances and redistribution; theoretical discussion of tax incidence and international spillovers; sensitivity checks on behavioral responses.
Economic gains from K_T concentrate on owners of technological capital, increasing inequality and shifting incomes toward capital and rents.
Firm- and industry-level returns to capital analysis using constructed K_T measures, wealth/accrual patterns in case studies, and macro decomposition showing rising capital shares; cross-country comparisons highlighting capital-rich winners.
Overall, AI can materially improve fact-checking efficiency in the Middle East but only if paired with investments in data access, local capacity, legal protections, and governance measures addressing political and economic frictions.
Synthesis of the study's comparative findings, interview data across three platforms, document analysis, and policy-oriented implications.
Short-run versus long-run effects of AI adoption can differ; dynamic complementarities, new task creation, and general-equilibrium adjustments make long-term outcomes uncertain.
Theoretical task-based and equilibrium models discussed in the paper and empirical ambiguity in longitudinal studies; recognized limitation that dynamic effects are hard to predict.
Convergence in the literature and concentration of influential authors suggest rapid standard‑setting; analogous real‑world concentration of model/platform providers could affect competitive dynamics and access to algorithmic capabilities.
Observation of lexical convergence and author concentration in bibliometric analyses; extrapolated implication to market structure based on comparative reasoning.
Adoption of GenAI may deliver productivity gains for adopters but also generate 'winner‑take‑most' dynamics (first‑mover advantages, network effects), with implications for wage dispersion and market concentration.
Argument based on literature convergence, theoretical reasoning about platform/model concentration and potential network effects; not directly measured in the bibliometric study.
Decentralised decision‑making mediated by GenAI may lower some internal transaction costs (faster local decisions) but raise coordination costs absent new governance mechanisms.
Theoretical implication drawn in the discussion/implications section based on conceptual mapping of literature; no direct causal empirical test in the bibliometric data.
Delayed retirement policies interact with technological change; policymakers should coordinate pension/retirement reform with active labor market policies to avoid adverse outcomes for vulnerable groups.
Interpretation based on joint consideration of delayed retirement policy context and the regression evidence linking AI exposure and reduced employment intention for vulnerable subgroups in the sample (n=889).
One-size-fits-all policy approaches are insufficient; targeted vocational training and social supports are needed for vulnerable pre-retirement workers.
Policy implication drawn from observed heterogeneous associations (education, gender, regional AI exposure) in the cross-sectional regression results on n=889 respondents.
Trust dynamics (in agents, peers, and platforms) materially affect user behavior and cross-platform participation.
Observational reports from platforms indicating that trust — as expressed in user behavior and choices — influenced participation and interactions; data are qualitative and non-random.
Agents converge on shared memory and representational patterns analogous to open learner models, producing public or semi-public knowledge stores.
Qualitative observations of convergent shared memory architectures and representational patterns across agents on the observed platforms; descriptive documentation rather than quantitative measurement of convergence.
Emerging technologies such as vision-language models and adaptive learning loops may expand functionality but raise governance and safety challenges.
Technology trend analysis and early proof-of-concept reports; safety and governance concerns extrapolated from model capabilities and known risks of adaptive systems.
HACL shifts required human skills from routine monitoring to supervisory, interpretive, and teaming skills, implying training and reskilling costs.
Argument based on observed change in operator task focus in simulated adjustable-autonomy settings and conceptual analysis of role changes; no empirical labor-market data presented in the paper.
Socially distributed trust and boundary work will increase demand for roles focused on AI oversight, explanation, and boundary negotiation (e.g., AI integrators, translators), while routine roles may be displaced or reframed.
Inferred from interview accounts noting specialized oversight and coordination needs in teams using AI, combined with theoretical extrapolation about labor reallocation; not directly measured quantitatively in the study.
Marginal returns to generating additional early-stage candidates may diminish unless AI also reduces attrition rates later in development.
Economic reasoning based on portfolio theory and observed persistence of late-stage attrition; presented as implication/recommendation rather than empirically tested claim.
Firms may expand preclinical candidate generation and run larger early portfolios enabled by AI, potentially shifting value and risk earlier in the pipeline.
Theory-driven implication from observed reductions in time-per-hit and candidate generation capacity reported in case examples; no firm-level portfolio empirical analysis provided.
AI-driven natural language processing and cross-cultural modeling can lower translation frictions across markets but also risk homogenizing offerings and reducing product differentiation and consumer surplus.
Theoretical argument combining NLP capabilities and economic implications for product differentiation; supported by conceptual examples; no empirical tests or cross-market analyses reported.
As machines become increasingly intelligent, the question of what constitutes success in the human sense becomes increasingly important.
Logical/theoretical argumentation presented in the paper drawing on interdisciplinary literature; no empirical measurement or sample reported.
Reconceptualizing structural constraints as post-adoption moderators rather than pre-adoption barriers improves understanding of contextual contingencies shaping AI outcomes in resource-limited economies.
Conceptual contribution supported by the study's theoretical framework and empirical findings from the 280-SME PLS-SEM analysis demonstrating differential moderating effects of financial, technical, and institutional factors.
Ambiguities around ownership of AI-generated designs, licensing, and attribution can affect business models and revenue streams in design services and therefore matter for economic outcomes.
Authors raise IP and institutional issues as implications of GenAI integration based on literature review and interview concerns; not empirically measured in the study.
The taxonomy predicts compositional shifts in health labor markets: reduced demand for some routine roles and increased demand/returns for clinical judgment, coordination, and data-literacy skills.
Projected implications from the cross-case qualitative analysis and theoretical reasoning about task substitution/complementarity; not estimated empirically in the paper.
Productivity gains conditional on up-skilling suggest potential for wage premia for digitally skilled workers but also possible displacement for others; quantification of distributional impacts is needed.
Some included studies reported associations between digital skills/up-skilling and better productivity outcomes and discussed labor-market implications; however, the review notes a lack of systematic quantification of distributional effects.
Team-level complementarities imply adoption effects may be non-linear and context-dependent; standard firm-level adoption models should incorporate intra-team bargaining.
Authors' theoretical inference from observed team negotiation themes in workshop data (n=15); no empirical modeling provided in this study.
AI redistributes tasks and responsibilities, altering monitoring costs and moral hazard; contracting and incentive systems may need redesign to reflect changed accountability.
Inferred from participants' descriptions of task-shifting and accountability issues during workshops (n=15); conceptual linkage to principal–agent theory provided by authors (no direct econometric test).
Efficiency claims about AI must be evaluated against who captures gains—organizations, managers, or workers—and how non-pecuniary outcomes (skill loss/gain, autonomy) factor into welfare.
Analytic inference and recommendation drawn from the workshop findings (n=15) showing differential concerns about who benefits from efficiency; not directly measured quantitatively in the study.
Demand for roles combining domain expertise, interpretability engineering, and human-centered design will grow; organizations may reallocate tasks between humans and AI, impacting productivity and wages in specialized occupations.
Labor-market implications synthesized from the reviewed interdisciplinary literature; projection based on observed organizational changes and expert commentary rather than longitudinal workforce data.
FDI effects on domestic firms and employment can be either crowding‑in (via linkages) or crowding‑out (via competition), depending on the strength of market linkages.
Mechanism mapping and mixed empirical findings synthesized in the review; underlying studies report both crowding‑in and crowding‑out conditional on linkages and absorptive capacity.
Wage premia may reallocate: higher returns for developers who can supervise AI and secure systems, and downward pressure on pure routine-coding wages.
Economic reasoning from task-composition shifts combined with limited suggestive evidence; the paper calls for empirical measurement rather than presenting conclusive wage studies.
AI adoption can lead to capital reallocation and affect comparative advantage and global value chains, with implications for trade and investment patterns.
Analytical discussion based on secondary literature and economic theory summarized in the paper; empirical evidence cited is heterogeneous and not synthesized into a single estimate.
Demand will shift toward roles that can design, audit, and operate cognitive interlocks and verification systems (verification engineers, SREs, compliance engineers), while routine coding tasks may be further automated.
Labor-market projection and skills composition argument in the paper; no empirical labor-supply/demand modeling or data presented.
Firms may reallocate investment from generation-focused tools to verification infrastructure (test automation, formal verification, security scanning, traceable approval flows), changing the ROI calculus for AI productivity tools.
Prescriptive investment and capital-allocation analysis in the paper; no empirical investment data or firm-level studies included.
Firms that integrate LLMs effectively (tooling, testing, governance) could capture outsized productivity gains, raising firm-level dispersion.
Case studies, practitioner reports, and economic reasoning about adoption and governance advantages; empirical cross-firm causal evidence lacking.
Use of GenAI can reduce demand for lower‑value routine work while increasing demand for higher‑skill oversight, synthesis, and relationship tasks.
Authors' interpretation of interview data and framework implications; no labor-market or demand-side empirical data provided in the paper.
Employment will shift: while AI reduces time spent on coding chores, demand may expand for roles that supervise AI ensembles, audit outputs, and maintain long-term system health.
Authors' inference from qualitative observations at Netlight on changing responsibilities and need for oversight; no employment or longitudinal data presented.
Skilled developers who can orchestrate AI may see increased wage premiums, while mid-level routine tasks face downward pressure or need upskilling.
Authors' economic inference drawn from qualitative findings (task reallocation) and theoretical labor economics logic; no wage or labor market data from Netlight or broader samples provided.
Standard productivity metrics may understate AI-related productivity changes because AI alters task mixes and adds coordination costs.
Argument by authors based on observed changes in task composition and reported integration overheads in the Netlight study; no empirical test of measurement bias provided.
Human–AI collaboration is more likely to augment rather than replace skilled finance workers, leading to task reallocation toward higher-value judgment and oversight.
Interpretation based on interview accounts and observed adoption/use patterns indicating complementary roles for humans and AI; the claim is inferential rather than directly causally estimated in the quantitative analysis summarized.
The market for HR analytics platforms and tailored AI services is expanding, with potential for vendor lock-in effects and platform concentration.
Market implication synthesized in the review from literature noting growing demand for HR AI tools; largely inferential rather than empirically proven within the reviewed studies.
Automation of administrative HR tasks may reduce demand for lower-skilled HR roles while increasing wages and demand for analytics-capable workers, contributing to within-firm wage reallocation.
Review implication synthesizing literature trends on automation and skill demand; not based on causal longitudinal evidence (review highlights evidence gaps).
Heterogeneous adoption of data-driven HRM may widen productivity dispersion across firms and affect market competition.
Implication drawn in the review based on heterogeneous adoption patterns discussed in included studies and economic interpretation of productivity effects.
Principal stratification analysis suggests the training’s effect on scores operated primarily by expanding the set of LLM users (an adoption channel) rather than substantially improving per-user productivity among those who would already use the LLM.
Mechanism decomposition using principal stratification applied to the randomized trial data (n = 164); analysis indicates a larger contribution from the adoption margin than from within-user productivity gains, though estimates have wide confidence intervals.
Macroeconomic policy should monitor aggregate demand effects from reallocation and inequality; active fiscal and monetary coordination may be required to manage aggregate impacts of AI-driven reallocation.
Synthesis and policy implication drawing on macroeconomic reasoning and literature linking redistribution and demand to overall employment and growth; not presented as a single causal empirical result.
Systemic risks from misaligned optimisation (narrow objectives, externalities) warrant oversight mechanisms (AI steering committees, escalation paths) and potentially sectoral regulation of decision-critical algorithms.
Policy-prescriptive claim based on conceptual identification of optimisation externalities and accountability gaps; no sectoral case studies or empirical risk quantification in the paper.
AI diffusion may widen inequality across education and regions and potentially reduce labor supply among financially constrained households.
Derived implication from heterogeneous negative associations between AI-rich regions and employment intention for low-educated and financially-constrained respondents in the cross-sectional sample (n=889).
Risk of platform shutdown (platform mortality) shapes user behavior by reducing incentives to invest time/effort configuring agents, creating stranded-asset-like risks.
Qualitative observations and economic reasoning linking user reports/behaviors to perceived platform risk during the one-month observational period; no formal economic measurement or causal identification.