Evidence (8807 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
Filtered →
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
Filter claims →
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 |
Productivity
Remove filter
Artificial intelligence industry agglomeration (AIIA) has a U-shaped relationship with agricultural pollution–carbon reduction synergy (APCRS) in the full sample.
Full-sample empirical analysis using panel regressions on data for 30 provinces (2016–2024) showing a nonlinear (U-shaped) estimated relationship between AIIA and APCRS.
The paper develops a task-to-firm conversion framework explaining why task-level GenAI productivity gains do not automatically translate into firm-level improvements.
Theoretical and conceptual contribution presented in the review, integrating multiple literatures (GPT theory, digital economics, task experiments, China studies).
Despite task-level gains, GenAI produces uneven or limited firm-level productivity effects in many settings.
Review synthesizing discrepancies between task-level experiments and firm-level outcome studies, and discussion of conversion frictions in the paper.
Generative AI (GenAI) should not be treated as a standalone productivity shock; its economic value depends on the interaction between model capability, task fit, human-AI calibration, organizational complementary assets, and regional digital infrastructure.
Conceptual framework developed in this review synthesizing literature from AI research, task-level productivity experiments, general-purpose technology theory, digital economics, and China-focused digital transformation studies; no new firm-level empirical analysis in this paper.
The study distinguishes foundational theoretical perspectives from the contemporary 2015–2025 evidence base and clarifies the relationship between task transformation and structural transformation, emphasizing institutional complementarity as the key mechanism shaping AI-driven growth outcomes.
Analytic separation of theoretical literature and empirical studies in the structured review (2015–2025); thematic mapping linking task-level changes to broader structural transformation contingent on institutional complementarities.
Rather than proposing a deterministic growth model, the study advances a conditional and ecosystem-centered interpretation of AI-led development.
Authors' interpretive conclusion based on their structured review and the integrative innovation-ecosystem framework synthesizing mechanisms and contextual dependencies in the 2015–2025 literature.
Interpreting task-based automation models alongside endogenous-growth and open-innovation frameworks clarifies why similar AI investments may lead to divergent structural outcomes.
Theoretical synthesis combining task-based automation literature with endogenous-growth and open-innovation models, illustrated by examples from the reviewed empirical literature (2015–2025).
The paper develops an integrative innovation-ecosystem framework linking three core transmission channels: (i) total factor productivity (TFP), (ii) task reallocation and labor-market restructuring, and (iii) innovation and knowledge-generation dynamics.
Conceptual framework constructed by the authors via integrative review of theoretical and empirical literature from 2015–2025; framework synthesizes mechanisms reported across studies.
Empirical evidence remains heterogeneous, and estimates of AI’s macroeconomic contribution vary across institutional and structural contexts.
Synthesis of heterogeneous empirical studies from the 2015–2025 literature identified in the structured review; comparative thematic classification highlighting variation by institutional/structural context.
AI adoption does not generate uniform or automatic growth effects.
Structured literature review / mechanism-oriented synthesis covering studies from 2015–2025; transparent search, screening and thematic classification (no formal meta-analysis).
The intended contribution is an Information Systems framework explaining when AI supports human augmentation and when it produces functional substitution.
Stated intended theoretical contribution in the abstract (proposed framework). This is an intended outcome rather than an empirically demonstrated result in the provided text.
The study investigates both perceived and enacted managerial agency.
Stated measurement targets in the abstract (descriptive of dependent variables). No measurement instruments or sample reported in the provided text.
The research uses a sequential multi-phase design combining experiments and qualitative fieldwork.
Stated methodology in the abstract (methodological claim about study design). No sample sizes or procedural details provided in the excerpt.
The study focuses on how technological design features, including transparency and override flexibility, interact with governance structures such as accountability and incentive systems.
Stated focus of the study in the abstract (descriptive of independent variables and governance moderators). No empirical details or sample reported in the provided text.
This doctoral research examines how AI-enabled decision systems affect human agency in data-driven organizations.
Stated research scope and aim in the paper (descriptive claim about the study's focus). No sample or results provided in the abstract.
Artificial intelligence is increasingly embedded in organizational decision-making, reshaping how managers exercise discretion and responsibility.
Stated as a background/motivation statement in the paper (literature-driven claim in the abstract). No empirical evidence or sample reported in the provided text.
Projected yield distributions vary substantially across locations, with some lower productivity sites exhibiting yield increases under future climate scenarios.
Results from simulated climate-projection experiments across multiple locations showing heterogenous yield distribution changes, including increases in some lower-productivity sites.
AI has a significant positive impact on value chain upgrading in the eastern and western regions of China, while its effect in the central region is insignificant.
Region-specific panel regressions / heterogeneity analysis using the 30-province 2010–2022 panel split by region; reported significance levels for eastern, western, and central subsamples.
There is a similar shift to agentic tooling outside OpenAI, particularly within organizations, although external adoption remains lower and more uneven.
Comparative usage analysis across three populations (external personal-account users, external organizational-account users, and OpenAI workers) from Codex logs.
Cluster analysis reveals diverse yet cohesive national profiles across the EU that reflect differences in digital readiness, human capital, and institutional factors.
Cluster analysis performed on country-level indicators (AI adoption, digital readiness, human capital measures, institutional factors) to group EU countries into profiles; summary reports heterogeneous but cohesive clusters; exact cluster counts and sample size not reported.
Some skills generalize broadly across tasks and models, whereas others become specialized to role-specific workflows and lose effectiveness under transfer.
Analyses reported in the paper showing heterogeneous transfer behavior across the 22 procedural skills in the AFTER benchmark, with some skills showing broad cross-task and cross-model generalization and others showing role-specific specialization and reduced transfer performance.
The Simpson's paradox in the pooled result is driven entirely by agent composition: Codex dominates 64.9% of the dataset.
Descriptive composition statistics from the AIDev dataset showing agent shares; explicit statement that Codex comprises 64.9% of dataset.
Better measurement matters, but improved measurement alone will not close the coordination gap between researchers and policymakers.
Authors' analytical conclusion arguing that measurement improvements are necessary but insufficient.
This article adopts a contextual approach to technology, considering it in conjunction with the social context in which it is situated.
Methodological statement made by the author about the approach taken in the paper (contextual rather than purely technical); not an empirical claim.
Longevity produces a short-run welfare loss that recedes as capital deepening raises wages, since households initially compress consumption and fertility to finance a longer retirement.
Model-derived welfare time path following a longevity shock showing initial welfare decline and subsequent recovery as aggregate capital deepens and wages rise; mechanism traced to household saving and fertility responses in simulations.
Robustness checks across the capital share, shock persistence, and the utility specification show that only an empirically implausible labor–AI elasticity reverses the wage and fertility signs.
Sensitivity/robustness analysis of model results by varying parameters (capital share, shock persistence, utility functional form) and the labor–AI elasticity, reporting conditions under which sign flips occur.
A forecast-error variance decomposition attributes most aggregate volatility to the longevity shock, while the AI shock dominates the variance of the return to AI capital.
Model-based forecast-error variance decomposition implemented on the simulated stochastic model to apportion variance of aggregate variables and the return to AI capital across shocks.
The two shocks move fertility in opposite directions: the AI shock raises fertility modestly through an income effect, while the longevity shock lowers fertility by strengthening life-cycle saving motives and increasing the cost of childrearing.
Endogenous-fertility overlapping-generations model with counterfactual simulations for AI and longevity shocks; comparative statics and simulation results regarding fertility responses and their mechanisms.
The AI shock reallocates investment from physical to AI capital.
Model simulation showing changes in investment allocation across capital types following the AI technology shock.
Stronger synchronization can increase collective output but may also increase systemic fragility and reduce mobility.
Analytical results and trade-off analysis in the model showing the effects of synchronization on collective output, fragility, and mobility; theoretical deduction without empirical sample.
The macroeconomic significance of AI-induced productivity depends not only on technological efficiency, but also on the distributive transmission of productivity gains through labour income, disposable income, prices, investment, public expenditure, transfers and external demand.
Theoretical argument and synthesis of literature in the conceptual review (no new empirical estimation reported).
AI-driven technological progress generates localized efficiency improvements while diffusing only weakly across the broader economy.
Synthesis of empirical results: localized positive associations between intangible capital and sectoral productivity versus weak/insignificant associations between AI patent intensity and aggregate TFP (analysis based on OECD Productivity, OECD STAN, INTAN-Invest, OECD Patents, FUAs; panel and robust regressions and descriptive work).
LLM guidance was associated with increased pupil size variability.
Physiological eye-tracking measure (pupil size variability) reported and compared across conditions in the simulated SAR experiment.
Eye-tracking data revealed an attention-guidance trade-off: visual resources shifted to the chat interface when LLM guidance was present.
Eye-tracking measures collected during the experiment showing changes in gaze allocation (increased fixations/dwell time on the chat interface) across LLM-guided vs baseline conditions.
The paper formalizes four mechanism theorems explaining the overhead-pressure dynamics: overhead non-additivity, augmentation-saved-time pathways, innovation-premium amplification, and human-AI dyad attribution uncertainty.
Presentation of four mechanism theorems within the paper (theoretical/mathematical exposition rather than direct empirical tests).
The ICH framework predicts three distinct augmentation regimes (determined by combinations of A and C) with distinct policy implications.
Theoretical classification derived from the model; conceptual prediction presented in the paper.
While AI has the potential to improve operational efficiency and strengthen adaptive capacity, inadequate readiness can increase systemic risks arising from algorithmic opacity, cybersecurity challenges, data dependence, coordination failures, and disruptions that may spread across interconnected administrative systems.
Conclusion drawn from the integrative conceptual framework and the systematic review of 68 empirical studies documenting both benefits and risks in different contexts.
Evidence on the productivity, risk, and resilience implications of AI adoption remains fragmented and dispersed across different fields of research.
Author's assessment of the literature based on the systematic review (PRISMA) of 68 empirical studies published 2015–2025.
Organisational performance becomes more dependent on the reliability of algorithms, the quality of data, effective governance, and coordination among public institutions.
Conceptual argument supported by synthesis of empirical studies in the systematic review (68 peer-reviewed empirical studies).
Artificial intelligence (AI) is becoming increasingly embedded in the digital infrastructure of local government, creating new opportunities to improve public sector productivity while also influencing systemic risk and organisational resilience across interconnected public systems.
Statement based on literature synthesis in the paper; theoretical framing and review of empirical studies (systematic review).
The endurance budget is dormant on premium 3,000-P/E TLC at datasheet prices and binding on the commodity QLC/eMMC (~1,000 P/E) that cheaper edge robots run.
Comparative statement based on device endurance specifications cited in the paper (3,000 P/E for TLC vs ~1,000 P/E for QLC/eMMC) and cost/pricing considerations; presented as boundary conditions for when the endurance budget matters. No empirical sample size reported.
Measured on real robot logs, the sign of the value-write association χ is a property of the deployment regime: positive on recurrent long-horizon manipulation (ĥχ ≈ +1.0 × 10^{-3}, replicated at full power), null on a shorter-horizon suite, and negative on non-recurrent teleoperation.
Empirical measurement on real robot logs at a pre-specified gate; reports an estimated value ĥχ ≈ +1.0 × 10^{-3} for recurrent long-horizon manipulation and qualitatively reports null and negative signs for other regimes. The paper states the +1.0e-3 estimate was replicated at full power. Exact sample size not reported in the excerpt.
The index is cost-optimal whatever the sign of the value-write association χ; only when χ > 0 does the optimum turn non-monotone, sending a robot's most valuable memories off its flash.
Theoretical result from the paper's model/analysis. The claim states a general optimality property (index cost-optimal for all χ) and a conditional structural result (non-monotone placement when χ>0). No empirical sample size reported.
The near-term value of Agentic AI does not lie in full autonomy or workforce reduction, but in controlled partial autonomy for simple and medium complexity business processes.
Central argumentative claim/recommendation in the paper (theoretical justification; no empirical study or sample size reported).
Quantile regression estimates reveal pronounced asymmetry across the biofuel production distribution: the AI effect is substantially stronger among low-production countries (Q10–Q25 elasticities: 0.58–0.61) and statistically insignificant among high-production countries.
Quantile regression analysis reported in the paper with elasticity estimates for Q10–Q25 and significance tests across quantiles.
The pattern of timing and magnitudes for publication volume and VC investment is theoretically consistent with a multi-stage technology diffusion process, implying two complementary pathways: a research output channel and a commercial adoption channel.
Interpretation based on differential lags and elasticities (2‑year lag for publications vs 1‑year for VC) and theoretical framing in discussion.
The effectiveness of AI in strategic core functions is contingent upon the human–AI interface.
Stated as a conditional claim in the paper—AI effectiveness depends on the quality of the human–AI interface; no empirical quantification provided in the summary.
Tranquil periods lower subjective risk assessments, raise AI substitution intensity, and compound leverage, generating a cognitive Minsky moment in which subjective risk falls while true systemic fragility rises.
Derived dynamics and comparative statics in the formal model; stated as one of the paper's propositions. No empirical data.
The transition is in trivia count, not rate; the gap 1-α is the unrecorded mass.
Analytic argument/proof in the model showing that whether trivia allowance is finite or infinite (count) determines the phase transition in achievable coverage, and identifying 1-α as the portion of valuable mass not recorded by the literature core.
Sharp dichotomy on the tight family: generators emitting finitely many trivia achieve optimal coverage α/2, while any infinite trivia allowance, even at vanishing rate, jumps the optimum to 1-α/2 (both tight, for cores presented as the candidate intersection), and one generator attains both ends.
Mathematical theorem(s) in the paper establishing tight upper/lower bounds on coverage for the 'tight family' under two regimes (finite trivia vs infinite trivia), expressed as functions of the core density parameter α.