Evidence (7560 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
Filtered →
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 |
Human Ai Collab
Remove filter
There will be accelerated demand for digital and specialised tech roles in India's IT sector by 2026.
Projection and analysis based on industry reports and workforce data (paper states it draws on industry reports and workforce data). Specific datasets, sample sizes, and statistical methods are not specified in the abstract.
In the digital economy, effective use of AI is crucial for maintaining supply chain stability in sports enterprises.
Argument supported by application of systems theory and supply chain management theory and substantiated by the paper's empirical results from the DML analysis of 45 listed Chinese SEs (2012–2023).
Talent attraction is the primary mechanism through which AI affects supply chain stability in sports enterprises.
Mechanism/mediation analysis within the DML framework applied to the 45-firm panel (2012–2023), showing talent attraction mediates the AI → SCS relationship more strongly than other tested channels.
The framework and roadmap offer actionable guidance for HRM practitioners, organizational leaders, and U.S. workforce policy stakeholders seeking to leverage AI for sustained competitive advantage.
Applied recommendations produced from the paper's conceptual synthesis; labeled as 'actionable guidance' in the summary (no outcome evaluation or pilot implementation results reported).
A balance between technological advancement and human capital investment is critical for minimising disruptions and ensuring a smooth transition to AI-driven operations.
Presented as a central conclusion from combining theoretical and empirical findings in the mixed-method study; the summary does not include quantification or sector-specific validation.
Organisations that integrate transparent governance and employee participation into AI adoption strategies experience lower resistance and higher acceptance.
Empirical insight reported by the study based on its theoretical analysis and Scopus-derived evidence; specific case studies are referenced but details (number of organisations, sectors, measures of resistance/acceptance) are not provided in the summary.
AI increases demand for advanced technical skills.
Reported as a main finding based on a mixed-method approach combining theoretical analysis and empirical insights from an analysis of records in the 'AI-driven transformation' Scopus database. (No sample size, statistical tests, or specific metrics provided in the summary.)
Federal funding for automation in specialty crops has been a focus of increased funding by both the US Department of Agriculture and the National Science Foundation, providing a path for innovators to produce automation and technology for nursery crops.
Statement in the paper about increased federal funding priorities (USDA and NSF); no specific program names, funding amounts, or timelines provided in the excerpt.
The percent of all tasks automated has increased approximately 15% over a 15-year period ending in 2021.
Comparison reported from a national labor survey (mid-2000s to 2021); exact survey methodology and sample size are not provided in the excerpt.
Use of the H-2A visa program has increased tremendously for the green industry in the past decade to help stop-gap the labor crisis.
Paper's statement about trend in H-2A program usage for the green industry; specific administrative data, years, or magnitudes not provided in the excerpt.
AI should be framed as augmentation rather than substitution, implying organizations need to invest in workforce upskilling in AI literacy to prevent harmful displacement and to enable designers to act as 'co-pilots' or 'AI curators'.
Interpretive and normative conclusion based on observed productivity/innovation benefits and literature/theoretical discussion; no firm-level employment displacement metrics reported in the study.
Managers should prioritize Generative Design and Predictive Analytics and adopt a 'Data-First' strategy (digitize historical assets and build digital infrastructure) to realize AI-enabled efficiency and innovation gains in design projects.
Managerial recommendations derived from the empirical findings linking AI to productivity and innovation gains; prescriptive guidance rather than empirically tested interventions within the paper.
AI functions as a bridge between project management efficiency and creativity in design projects, enabling automation of routine workflows and freeing designers to focus on higher-value creative tasks.
Interpretation based on empirical findings (AI positively associated with TFP and innovation) and mechanism discussion; supported by text-analysis results and conceptual framing in the paper (no granular project-level workflow logs presented).
LLMs are increasingly supporting decision-making across high-stakes domains, requiring critical reflection on the socio-technical factors that shape how humans and LLMs are assigned roles and interact during human-in-the-loop decision-making.
Background/positioning claim supported by cited literature and the authors' motivation for the work (trend observation). Specific empirical support is not detailed in the abstract.
Technological progress has historically contributed to productivity and economic growth.
Asserted in the paper as a historical generalization within the conceptual analysis; no original empirical data or sample provided in this paper to quantify the effect.
The integration of Fuzzy BWM-PROMETHEE II-DEMATEL framework constitutes a novel methodological contribution and provides useful decision support for strategic planning and resource allocation.
Authors' methodological claim in the paper that combining these fuzzy MCDM techniques is novel and yields decision-support outputs; novelty and practical utility asserted but not externally validated in the provided summary.
Addressing High Initial Investment and Supply Chain Integration initially helps accelerate digital readiness and enhance transformation performance.
Inference/recommendation derived from the PROMETHEE II and DEMATEL results that mark these two factors as dominant causal drivers; no reported empirical intervention or longitudinal validation in the provided text.
Fuzzy BWM results highlight Customization, Flexible Production, Human–Machine Collaboration, and Cybersecurity as the most influential practices supporting I4.0 implementation.
Results reported from the paper's Fuzzy BWM analysis informed by literature survey and expert judgments. (Exact number/composition of experts and statistical details not provided in the supplied summary.)
AI tools (e.g., machine learning) can mitigate managers' information processing constraints and thereby help integrate tax planning with core business strategy (i.e., AI mitigates constraints to improve effective tax planning).
Inference from observed positive relationships between AI investment and tax effectiveness, stronger effects in complex/high-status-tax-function firms, and mediation through information quality and capital management in the 2010–2018 U.S. nontechnology firm sample.
AI improves tax effectiveness by enhancing internal capital management.
Mechanism/mediation tests showing AI investment is associated with improved internal capital management, which is associated with increased tax effectiveness.
AI improves tax effectiveness by enhancing internal information quality.
Mechanism/mediation tests in the empirical analysis showing AI investment is associated with higher internal information quality, which in turn is associated with greater tax effectiveness.
The positive association between AI investment and tax effectiveness is concentrated among firms where the tax function holds a higher status.
Moderation analysis in the sample indicating the AI–tax effectiveness effect is stronger when the tax function has higher organizational status.
The positive association between AI investment and tax effectiveness is concentrated among more complex firms.
Subgroup/moderation analyses in the same sample (U.S. nontechnology firms, 2010–2018) showing stronger AI–tax effectiveness relationship for firms classified as more complex.
Investment in AI-related human capital is positively associated with tax effectiveness.
Empirical analysis using a recently developed firm-year measure of investment in AI-related human capital for a broad sample of U.S. nontechnology firms between 2010 and 2018; reported positive association (regression-based).
AI features increase consumers' purchase intention for electronic products.
Reported relationship tested via SEM between AI feature constructs and the dependent variable 'purchase intention' in the structured questionnaire data (no sample size or statistical details provided in the summary).
AI features improve perceived decision-making support (i.e., ease of decision-making / simplification of product evaluation).
Reported empirical result from SEM analysis of questionnaire data measuring perceived decision-making support as a dependent variable.
AI features positively affect consumer trust.
Empirical finding reported from SEM analysis linking AI features to the dependent variable 'consumer trust' in the questionnaire data (specific effect sizes/p-values not provided in the summary).
AI-enabled features significantly enhance consumer confidence and satisfaction by simplifying product evaluations and increasing perceived usefulness.
Reported empirical result from this study based on a quantitative research design using a structured questionnaire and analyzed with Structural Equation Modeling (SEM). (Sample size and specific statistical values not reported in the summary.)
The SDK provides interoperability via MCP and A2A.
Implementation and interoperability description in the paper claiming MCP and A2A support; can be verified in code and integration tests.
AESP enforces the invariant that agents are economically capable but never economically sovereign.
Formal design of the protocol and five enumerated mechanisms described in the paper (policy engine, human review, EIP-712 commitments, HKDF isolation, ACE-GF substrate). Enforcement claim derives from architectural guarantees rather than empirical validation in the abstract.
The Agent Economic Sovereignty Protocol (AESP) is a layered protocol that lets agents transact autonomously at machine speed on crypto-native infrastructure while remaining cryptographically bound to human-defined governance boundaries.
Protocol design and specification presented in the paper; implementation claimed (see TypeScript SDK). No runtime throughput/latency measurements reported in the abstract.
Rigorous user evaluation can help develop systems that allow for effective and responsible human agency in veracity-assessment processes.
Interpretation and conclusion drawn from the study's findings showing differences in user behavior across support types and highlighting design implications for tooling that supports human verification.
User responsibility for assessing veracity is particularly critical in sectors that rely on on-demand, AI-generated data extraction, such as biomedical research and the legal sector.
Framing and motivation in the paper (domain argument citing the high-stakes nature of biomedical and legal information extraction). This is a contextual claim rather than a new empirical result from the study.
LLM explanations enable rapid veracity assessments.
Same controlled user study, where assessment speed (time-to-decision) was measured for the LLM-explanation condition and found to be fast relative to other conditions.
Passage retrieval offers a reasonable compromise between accuracy and speed, with judgments of veracity comparable to using the full source text.
Controlled user study comparing three types of supporting information (full source text, passage retrieval, LLM explanations). Outcome measures reported include veracity-judgment accuracy and time-to-assessment. (Sample size and statistical details not specified in the abstract; see paper for exact n and tests.)
Tailoring AI explanations to individual users can improve human–AI team performance and provides insights into how personalization may enhance human-AI collaboration.
Synthesis of experimental findings across the two preregistered tasks: observed interactions between user characteristics and explanation types, and demonstration of complementarity in the geography task, form the basis for this general claim. (This is an inferential conclusion drawn from the experiments; full generalizability depends on task scope and replication.)
In the geography-guessing task, user characteristics interact with explanation types, and these interactions contribute to human–AI complementarity (the joint performance exceeds either alone).
Results from the preregistered geography-guessing experiment showing interaction effects between user characteristics and explanation types that lead to observed complementarity. (Exact effect sizes, statistical significance, and sample size not provided in the excerpt.)
We designed a geography-guessing task in which humans and AI possess complementary strengths.
Task design described in the paper intended to generate complementary error patterns between humans and the AI model (methodological claim based on experimental design). (Details on design specifics and validation not provided in the excerpt.)
AI has the potential to deliver predictive benefits for recruitment and retention.
Aggregated findings from empirical studies in the systematic review and supporting meta-analytic/qualitative evidence across the 85 publications that examine recruitment/retention applications.
The meta-analysis shows a small-to-moderate direct positive relationship between AI use and operational productivity (r = 0.28).
Quantitative meta-analysis reported in the paper; pooled effect size r = 0.28; heterogeneity I^2 = 74% (based on the meta-analytic sample drawn from the reviewed studies).
Generative AI (GenAI) offers transformative potential for productivity and innovation.
Synthesis of themes reported across the 28 reviewed papers (authors' thematic summary of literature highlighting potential productivity and innovation gains).
Short-term productivity gains are documented.
Findings from some of the 81 reviewed sources report short-term productivity improvements associated with Agentic AI or related interventions. The abstract does not quantify the gains or specify domains/settings.
Analytics can serve as the focal interpretive intercession between AI outputs and human decision-makers, facilitating transparency, accountability, and contextual decision-making.
Conceptual proposition drawn from interdisciplinary literature synthesis and the proposed framework. No empirical validation or measured outcomes presented.
The workforce should be prepared for GenAI-driven changes through targeted skilling programs (upskilling, reskilling, cross-skilling).
Recommendation based on literature and the authors' analyses/discussions; no trial data or program evaluation metrics are reported in the abstract.
Using suitable approaches to skill development and committing to continuous learning within organizations, GenAI drives innovation, improves decision-making, and creates new growth opportunities.
Conclusion drawn from the paper's literature recherche, task analyses (including Erasmus+ projects), and discussions with trainers/educators. The abstract does not present controlled empirical evidence or quantified effect sizes for these outcomes.
GenAI supports skill-assessment tools that enable continuous, granular evaluations of employees’ abilities.
Supported by literature synthesis, analysis of occupational tasks (Erasmus+ projects), and practitioner discussions; no quantitative validation (e.g., accuracy, reliability, sample sizes) reported in the abstract.
GenAI supports learning and development by performing various tasks that influence the creation and interaction with content.
Claim based on reviewed literature and task analyses presented in the paper; specifics of experiments or deployment (e.g., tools used, participant counts) are not provided in the abstract.
Upskilling, reskilling, cross-skilling, and learning initiatives are necessary mechanisms for organizations to prepare their workforce for GenAI-driven changes.
Derived from literature recherche and analysis of individual tasks across occupations within Erasmus+ projects, plus practitioner discussions; no sample sizes or outcome metrics specified.
Generative AI (GenAI) models are growing rapidly, changing job roles, and revolutionizing entire industries.
Stated by the authors based on a literature recherche (scope and search strategy not specified in abstract). No quantitative sample size or bibliometric details provided.
LLM use increases information overload (additional analyses).
Reported follow-up/additional analyses from the experiment indicating a statistically significant association between LLM use condition and higher scores on information-overload measures.