Evidence (871 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
20058 claims
Filter claims →
Productivity
17184 claims
Filter claims →
Governance
16099 claims
Filter claims →
Human-AI Collaboration
16034 claims
Filter claims →
Innovation
10501 claims
Filter claims →
Org Design
10496 claims
Filter claims →
Labor Markets
6444 claims
Filter claims →
Skills & Training
5385 claims
Filter claims →
Inequality
4148 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 | 1820 | 479 | 278 | 1820 | 4588 |
| Organizational Efficiency | 2711 | 616 | 401 | 173 | 3922 |
| Governance & Regulation | 2075 | 886 | 459 | 246 | 3714 |
| Technology Adoption Rate | 1467 | 530 | 258 | 206 | 2488 |
| Decision Quality | 1281 | 496 | 289 | 152 | 2228 |
| Output Quality | 1227 | 447 | 207 | 138 | 2025 |
| AI Safety & Ethics | 634 | 754 | 207 | 83 | 1688 |
| Research Productivity | 826 | 241 | 114 | 422 | 1624 |
| Firm Productivity | 1052 | 154 | 163 | 66 | 1441 |
| Task Allocation | 685 | 211 | 331 | 99 | 1335 |
| Market Structure | 433 | 423 | 242 | 46 | 1150 |
| Innovation Output | 639 | 91 | 105 | 34 | 871 |
| Task Completion Time | 476 | 113 | 43 | 36 | 672 |
| Firm Revenue | 445 | 126 | 58 | 25 | 656 |
| Skill Acquisition | 364 | 119 | 109 | 34 | 626 |
| Consumer Welfare | 288 | 167 | 104 | 31 | 592 |
| Employment Level | 214 | 140 | 174 | 50 | 582 |
| Error Rate | 230 | 251 | 35 | 16 | 535 |
| Fiscal & Macroeconomic | 268 | 136 | 71 | 50 | 532 |
| Inequality Measures | 100 | 307 | 96 | 12 | 515 |
| Worker Satisfaction | 221 | 173 | 60 | 30 | 484 |
| Automation Exposure | 155 | 138 | 65 | 36 | 398 |
| Regulatory Compliance | 171 | 120 | 30 | 13 | 335 |
| Developer Productivity | 222 | 58 | 27 | 13 | 321 |
| Team Performance | 188 | 56 | 50 | 24 | 320 |
| Wages & Compensation | 146 | 104 | 46 | 16 | 312 |
| Training Effectiveness | 207 | 41 | 21 | 26 | 298 |
| Job Displacement | 23 | 153 | 52 | 4 | 232 |
| Hiring & Recruitment | 102 | 57 | 30 | 11 | 202 |
| Skill Obsolescence | 16 | 102 | 24 | 6 | 148 |
| Creative Output | 71 | 42 | 23 | 6 | 143 |
| Social Protection | 57 | 30 | 11 | 3 | 101 |
| Labor Share of Income | 29 | 42 | 24 | 2 | 97 |
| Worker Turnover | 43 | 29 | 6 | 4 | 82 |
| Industry | — | — | — | 1 | 1 |
Science has a positive local effect on co-located digital and artistic activity while exerting a negative regional backwash effect that draws creative capacity away from neighbouring districts.
Spatial Durbin model estimating within-district and between-district effects for digital technology, science, and arts segments.
Creative agglomeration in Slovakia is conditional on reaching segment-specific density thresholds and is shaped by asymmetric cross-district spillovers and industrial legacy.
District-level analysis using segment-specific location quotients, population density, manufacturing specialization, a spatial Durbin model, random-forest simulations, and breakpoint tests.
Governance-by-design may shift innovation toward safer and more auditable systems, while potentially slowing time-to-market.
Theoretical and interpretive analysis of how embedded compliance requirements may affect product design and innovation incentives; no quantitative innovation or time-to-market estimate is provided.
Digital-green synergy is more strongly associated with strategic, breakthrough, and exploitative innovation than with substantive, incremental, and exploratory innovation.
The study compares associations across different innovation types using panel-level empirical analysis.
Data volume alone does not guarantee innovation; data quality and effective use are more important for realizing innovation value.
Review synthesis citing Ghasemaghaei and Calic (2020), which is listed in the evidence table as a firm-innovation study.
The magnitude of the AI–green technology innovation relationship depends on local economic, industrial, and innovation conditions rather than being uniform across cities.
Distributional estimates from panel quantile regression and heterogeneity analyses by city size, resource dependence, and regional location.
The strength of the positive AI–green technology innovation relationship varies by city size, resource dependence, and regional location; it is relatively stronger in large cities, more stable in non-resource-rich cities, and strongest in western China.
Heterogeneity analysis stratifying the city-level panel by city size, resource dependence, and geographic region.
Benchmark incentives structure the AI community's innovation efforts by motivating researchers to design systems that compete well on established benchmarks.
Conceptual argument supported by citations to Orr and Kang (2024); no original causal or quantitative analysis is reported.
Generative AI accelerates creative content production and cultural circulation while altering the returns to different forms of DFC.
Theoretical synthesis linking generative AI, creative labor, and digital-capital returns; no large-scale quantitative causal test is reported.
Heightened security concerns are expected to accelerate dual-use AI research and increase tensions between broad diffusion for economic growth and controls intended to manage security risks.
Forward-looking analysis of dual-use research incentives, export controls, vetting regimes, and provenance tracing; no empirical estimates of research growth or control effects were provided.
Neither unrestricted scientific openness nor wholesale technological decoupling is adequate: unrestricted openness creates security risks, while decoupling can stifle innovation.
Comparative conceptual reasoning contrasting the risks and costs of the two policy extremes; not supported by a new quantitative estimate.
Ex ante constraints may reduce incumbents' ability to use platform rents to subsidize risky AI research and development, potentially slowing some large-scale innovation, while also reducing anti-competitive foreclosure that can hinder startup innovation.
Trade-off analysis based on the legal and economic implications of ex ante platform regulation; no empirical estimate is provided.
The paper's Digitally Enabled Fragmentation framework defines a condition in which digital technologies sustain global connectivity while geopolitical and institutional divisions constrain cross-border innovation activities.
Conceptual framework developed through structured literature review, open coding, axial coding, and theoretical integration across three analytical levels.
The positive relationship between AI application and green technological innovation operates partly through R&D investment, financing constraints, and managerial overconfidence.
The chapter's abstract and mechanism-analysis description identify these three channels; mechanism tests use firm-level fixed-effects regressions with differing available observations.
The appearance of predicted relationships in later literature validates knowledge association emergence but does not establish technological feasibility, scalability, or commercial value.
The study uses later-literature emergence as temporal validation, while explicitly identifying the absence of direct evidence on commercial viability and technical scalability as a limitation.
The patent landscape reveals divergent geographical, actor, sectoral, and functional orientations rather than a single clear direction of the AI-enabled energy transition.
The study applies an anticipatory framework that classifies patenting across four dimensions: geographical, actor, sectoral, and functional.
The jurisdictions differ in the types of actors responsible for patenting: private corporations dominate patenting in Europe, Japan, and the United States, whereas universities and public institutions play a larger role in China and South Korea.
Classification of patent applicants by actor type across the five jurisdictions in the patent-landscape analysis.
The moderating effect of digital transformation is concentrated in examination-intensive green invention patents and is absent for utility model registrations.
Outcome disaggregation comparing examination-intensive invention patents, which receive greater scrutiny, with utility model registrations.
The study measures formal inter-city innovation collaboration using joint patent applications and does not directly observe open-source, platform-based, remote, consulting, or informal knowledge-exchange collaboration.
The paper defines its empirical outcome using city-pair patent data and explicitly identifies the unobservability of non-patent collaboration channels as a limitation.
In the model’s new long-run equilibrium, innovative effort is higher but the productivity gains from innovation are lower because design longevity has declined.
The paper compares transitional and equilibrium outcomes after introducing a permanent increase in the obsolescence rate.
Two standard models of data & artificial intelligence innovation project management were identified, which demonstrated various outputs in the studied institutional contexts.
Descriptive classification/typology produced in the study comparing two management models and their observed outputs across institutional contexts (methods and sample not specified in excerpt).
Two standard models of data & artificial intelligence innovation project management were identified, which demonstrated various outputs in the studied institutional contexts.
Comparative classification/analysis reported in the paper identifying two project-management models and observing differing outputs across institutional/cluster contexts; methodological details and sample size not provided in the excerpt.
Digital platform concentration can either promote or hinder green innovation and sustainable development.
Conceptual analysis linking platform concentration to potential positive and negative effects on green innovation; literature-informed discussion without empirical quantification.
There is substantial variation of model efficiency within companies; a firm can train two models with more than 40x compute efficiency difference.
Within-firm comparisons of models in the dataset showing instances where two models from the same developer differ in compute efficiency by a factor greater than 40.
The importance of developer-specific advantages depends on where models lie in the performance distribution.
Heterogeneous effect estimates across performance quantiles/regions of the performance distribution reported in regression analysis.
AI influences agricultural pollutant emissions through green innovation capacity: it initially boosts but later weakens firms’ green innovation capacity.
Mechanism analysis reported in the paper based on firm-level empirical work (semi-parametric additive model); the authors report a nonlinear effect of AI on green innovation capacity (first positive then negative), affecting emissions outcomes.
Generative AI is changing how software is produced and used.
Stated as background premise in the paper (abstract). No empirical data reported in the abstract.
The effect of AI washing on the technological gap is also mediated by industry-level R&D investment.
Mediation analysis showing industry R&D intensity channels part of the AI-washing → technological gap relationship using panel data of China's A-share firms (2007–2022).
The effect of AI washing on the technological gap is mediated by firm-level R&D investment.
Mediation analysis conducted on the same A-share firm panel (2007–2022) showing internal R&D investment transmits part of the AI-washing effect to the technological gap.
AI washing initially widens the technological gap (stifles innovation) at low levels, but after a threshold it narrows the gap as firms adapt.
Interpretation of the inverted U-shaped regression results from the panel of China's A-share listed firms (2007–2022); reported sign change across AI-washing intensity.
There is a significant inverted U-shaped relationship between 'AI washing' and the corporate technological gap.
Empirical analysis using panel data of China's A-share listed firms (2007–2022); regression models testing nonlinear (quadratic) relationship.
Invention alone is insufficient to create game-changing businesses; real transformation requires alignment of technological advances with business models, organizational capabilities, ecosystem coordination, and societal purpose.
Argument supported by multiple qualitative case studies from sectors such as cinema, digital platforms, entertainment, energy systems, and AI described in the chapter.
While U.S.-led semiconductor export controls reliably generate short-term disruption and frontier delay, they tend to narrow rather than widen long-term capability gaps unless paired with sustained multilateral coordination and domestic innovation reinforcement.
Synthesis of the four-indicator analytical framework applied to qualitative data and comparative policy responses from 2019–2025; argument based on observed short-term impacts and longer-term adaptation patterns in case studies.
Sustained denial pressure systemically reallocates and intensifies innovation effort, accelerating domestic investment, architectural redesign, and the formation of parallel innovation pathways.
Comparative analyses and case studies showing shifts in R&D focus, reported increases in domestic investment and architectural redesign efforts in response to export controls (qualitative evidence, 2019–2025 period).
While IPRs incentivize innovation, they also generate monopolistic power.
Theoretical argument and literature review contrasting incentive effects of IPRs with their monopolistic consequences; no reported empirical sample size.
Artificial intelligence is fundamentally transforming Nepal's socio-economic landscape, with far-reaching implications across diverse sectors.
High-level assertion in the paper summarizing perceived impact; supported by the paper's sectoral analyses but no quantitative national-level metrics reported in the excerpt.
Platform laborers play an indispensable yet hidden role in building and sustaining AI systems.
Eight-month ethnography of Bangladesh's platform labor industry; qualitative observations and interviews (duration reported; sample size not stated in abstract).
The external institutional environment (notably government R&D subsidies and intellectual property protection) significantly moderates these relationships: it generally weakens (diminishes) the positive effect of centrality but strengthens the positive effect of structural holes during the rapid expansion phase.
Moderation tests incorporating measures of government R&D subsidies and intellectual property protection interacting with centrality and structural-hole variables in models of regional innovation performance using provincial patent data (2010–2024).
There is an inverted U-shaped relationship between structural holes (brokerage positions) in the network and regional collaborative innovation performance.
Regression analyses using structural-hole measures derived from provincial patent-collaboration networks (2010–2024) showing a quadratic association with regional innovation performance.
There is an inverted U-shaped relationship between network centrality and regional collaborative innovation performance across different developmental stages.
Statistical analysis relating provincial centrality measures (from SNA on patent co-application networks, 2010–2024) to regional collaborative innovation performance, reported as an inverted U-shaped (quadratic) effect.
Comparative analysis shows China's move toward self-reliant AI stacks (e.g., Huawei Ascend chips) and policy-driven innovation, contrasting with U.S. capital-intensive models.
Comparative synthesis using external benchmarks/reports (MERICS, RAND) and the authors' analysis; examples cited (Huawei Ascend).
The impact of the NAIPZ/DIT policy on corporate GI is heterogeneous and is moderated by firm size and industry characteristics.
Heterogeneity analysis reported in the paper showing differential policy effects across firms of different sizes and across industry types.
Generative AI is transforming work, creativity, and economic security in ways that extend beyond automation and productivity.
Conceptual framing supported by the paper's hybrid approach (labour market modelling, sectoral diffusion analysis, policy review, qualitative critique); no empirical sample size reported for this general claim.
Innovation capability and logistics efficiency appear to function as complementary mechanisms linking perceived AI utilization to perceived firm performance, with a smaller capability-to-process (innovation→efficiency) pathway.
Interpretation based on the pattern of significant individual and sequential indirect effects from the mediation analyses on the n=254 cross-sectional survey; authors characterize the mediators as complementary with a smaller capability-to-process path.
The research contributes by connecting AI adoption to inclusive economic modernization and proposing a governance-based framework for managing its risks in low- and middle-income contexts.
Originality / Value section claims conceptual contribution and a proposed governance framework; based on the paper's synthesis of comparative data and theoretical discussion (not an empirically validated framework in this study).
There are two distinct regional catch-up trajectories: Digital Leapfrogging in the Baltic States and Industrial Deepening in the Visegrad Group.
Systematic empirical documentation across the Visegrad Group and Baltic States (2022–2025) using the paper's assessment approach; patterns labeled and interpreted by the author.
The LCCP effect on AI industry development varies across local contexts, with stronger effects observed in established innovation hubs and in some follower regions undergoing industrial transition.
Heterogeneity analyses in the staggered DID framework on the 285-city panel (2007–2022) that split the sample by city type/region (innovation hubs vs. followers/industrial-transition regions) and report differential policy coefficients.
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.
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.