Evidence (9875 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
Filtered →
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
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 |
Adoption
Remove filter
The main empirical findings are robust to alternative model specifications and checks.
Paper reports robustness checks (alternative control sets, specifications, and sensitivity analyses) in which the negative IR–IWE relationship remains qualitatively unchanged.
Policy implication: AI functions as a complement to digital trade, increasing local economic and housing-market returns to digitalization; therefore, AI investments can be targeted to help lagging (non-coastal, low-income) cities capture benefits of digital trade.
Inference drawn from the positive moderation effect of the urban AI index on the digital-trade → house-price relationship and the stronger AI-driven effects reported for non-coastal and low-income cities.
AI adoption markedly increases the impact of digital trade on house prices in non-coastal and low-income cities, implying scope for digital catch-up.
Subgroup analyses and interaction estimates showing a stronger positive moderation effect of the urban AI index in non-coastal and low-income city subsamples (specific estimates and significance not provided in the summary).
Digital-trade effects on house prices are larger in high-income cities than in low-income cities.
Heterogeneity analysis by city income groups (high- vs low-income); reported stronger digital-trade coefficients in high-income cities (details of income cutoffs and sample sizes not specified).
Digital-trade effects on house prices are larger in coastal cities than in non-coastal cities.
Heterogeneity analysis splitting the sample by coastal versus non-coastal cities; reported stronger coefficients for coastal cities (specific sample counts and coefficients not provided).
Urban AI adoption positively moderates the effect of digital trade on city-level house prices: cities with higher AI capability experience a larger house-price response to digital trade.
Interaction terms in city-level panel regressions between the digital trade index and an urban AI index constructed via text-mining. Heterogeneity/interaction estimates reported (specific coefficients and significance levels not provided in the summary).
Recommendation: support capacity building—digital literacy, agronomic knowledge, and extension systems—to increase adoption and equitable benefits.
Authors' recommendation derived from recurring findings on human-capacity constraints in the reviewed studies.
AI interventions supported economic transformation in some contexts by improving market access and enabling reallocation toward higher-value tasks.
Findings from selected studies and institutional reports documenting improved market linkages, price discovery, and shifts in farm household activities.
AI applications contributed to environmental resilience via water and fertiliser savings and earlier pest detection in some studies.
Reported resource-use metrics and earlier detection outcomes in several reviewed studies and case reports synthesized thematically.
AI-enabled interventions produced technical efficiency gains through better input targeting and reduced waste.
Studies in the review reporting improvements in input targeting (e.g., fertiliser/pesticide application) and reductions in waste; aggregated in thematic synthesis.
AI deployment has produced measurable supply-chain efficiency improvements and better market integration in reviewed cases.
Synthesis of studies and institutional reports reporting metrics/qualitative evidence on logistics, aggregation, price discovery, and market linkages.
AI interventions are associated with input cost reductions up to ~25%.
Comparative effect-size synthesis across reviewed studies reporting input cost outcomes (2020–2025).
Across reviewed studies (2020–2025), AI interventions are associated with yield gains of roughly 12–45%.
Comparative effect-size synthesis of reported impacts across the reviewed studies (>60 articles/reports) that reported yield outcomes.
AI-powered digital agriculture in developing contexts—especially Sub-Saharan Africa—can materially improve productivity, sustainability, and rural livelihoods.
Structured literature review and thematic synthesis of >60 peer-reviewed articles and institutional reports (timeframe 2020–2025) focused primarily on Sub-Saharan Africa and other developing contexts.
Standards and open interoperability reduce vendor lock‑in and transaction costs, widening market access and competition for AI services built on DT data.
Economic reasoning and thematic findings from the literature linking interoperability to reduced transaction costs and broader market participation.
Public procurement and large asset owners can act as demand‑pulls to de‑risk early investment and help set standards for DT adoption.
Policy recommendation and examples from literature arguing that large buyers can catalyse adoption; based on case/policy studies in the review.
Better data continuity across lifecycle phases reduces model training friction and increases the value of historical data for forecasting and causal analysis.
Conceptual argument supported by case evidence in the review showing fragmented data reduces reusability; authors infer benefits for AI training and forecasting.
DTs generate continuous, high‑resolution operational data (IoT telemetry, usage patterns, maintenance logs) that can substantially improve AI models for predictive maintenance, scheduling, energy optimisation, and logistics.
Logical implication and examples from pilot studies in the review showing richer telemetry and operational datasets produced by DT pilots; argued benefits for AI model inputs.
Three core differences by which DTs extend BIM: (1) bidirectional automated physical↔digital data exchange; (2) integration of heterogeneous, real‑time sources (IoT, operational systems); (3) lifecycle continuity preserving data across handovers.
Conceptual synthesis across the literature reviewed (conceptual papers, case studies, pilots) identifying functional distinctions between DT and BIM.
Digital twin (DT) technology can materially improve construction lifecycle performance beyond what Building Information Modelling (BIM) delivers.
Synthesis of 160 reviewed studies including conceptual papers, case studies and pilot deployments reporting performance improvements attributed to DT implementations.
ANN analysis ranks information barriers as the most important predictor of organizational inertia.
ANN feature-importance analysis reported in the paper that ranks predictors for inertia, identifying information barriers as the top predictor; methodological specifics (sample size, ANN parameters) are not provided in the abstract.
Artificial neural network (ANN) analysis ranks functional values as the most important predictor of initial trust.
ANN feature-importance analysis reported in the paper that ranks predictors for initial trust, with functional values highest; method described as ANN-based relative importance ranking (details such as network architecture, training sample size, or validation metrics not reported in the abstract).
Human interaction, information, and norm barriers increase organizational inertia (resistance to change) toward GAICS.
Qualitative phase surfaced these barriers; quantitative validation showed statistically significant positive relationships between (a) need for human interaction barriers, (b) information barriers (lack of knowledge/clarity), and (c) norm barriers (cultural/social norms) and organizational inertia.
Functional and instrumental values increase initial trust in GAICS.
Mixed-methods evidence: qualitative exploratory phase identified functional and instrumental value as drivers; quantitative phase (inferential analysis) found positive, statistically significant effects of functional value (system usefulness/quality) and instrumental value (task-related benefits) on initial trust.
AI/ML–based credit scoring and alternative‑data underwriting reduce information asymmetries, lowering search and monitoring costs and expanding effective credit supply to previously rejected MSMEs and startups.
Analytical argument supported by illustrative case examples and literature on machine‑learning underwriting; the paper notes limited causal identification and time‑sensitivity of fintech products.
Government action (digital ID, payments rails, credit guarantees, standards, consumer protection) is vital to enable beneficial outcomes from digital finance for MSMEs.
Policy synthesis and comparative evaluation recommending government infrastructure and regulatory measures; conclusion based on institutional analysis rather than experimental evidence.
Case studies indicate FinTech platforms have meaningfully lowered rejection rates and loan turnaround times for underbanked MSMEs, accelerating working‑capital access.
Illustrative case studies of FinTech deployments in India reporting lower rejection rates and faster approvals; paper explicitly notes these cases are illustrative and not nationally representative and do not establish causal identification.
Supply‑chain financing can meaningfully unlock working capital for MSMEs by leveraging buyer creditworthiness, yielding high impact for MSMEs embedded in modern supply chains.
Comparative evaluation and illustrative case studies highlighting supply‑chain finance deployments; evidence is demonstrative and not nationally representative or causally identified.
Optimal financing outcomes generally come from hybrid approaches that combine formal banking credibility and policy support with FinTech speed and data-driven underwriting.
Comparative evaluation and policy synthesis recommending co‑lending, credit guarantees, and partnerships (banks as liquidity providers combined with FinTech underwriting); based on qualitative tradeoff analysis rather than experimental/causal evidence.
Compared with traditional bank loans and government schemes, contemporary financing models tend to be faster, more flexible, and more scalable for smaller firms.
Comparative qualitative evaluation across five variables and illustrative case studies showing reduced loan turnaround times and improved accessibility for small firms; no nationally representative sample or causal inference provided.
Digital technologies — especially FinTech lending platforms, alternative debt/equity products, supply‑chain finance, crowdfunding, and emerging blockchain applications — are materially expanding timely access to capital for Indian MSMEs and startups.
Multi‑criteria comparative evaluation (accessibility, finance cost, flexibility, risk, scalability) plus illustrative case studies of FinTech and alternative financing deployments in India that report faster turnaround and inclusion effects. The paper notes case evidence is illustrative rather than nationally representative and lacks quantitative causal identification.
Regulating algorithmic transparency, data practices, and truthful sustainability claims is important to preserve digital trust and efficient market outcomes.
Policy recommendations and economic reasoning presented in the paper; grounded in literature on algorithmic governance and consumer trust; not empirically validated within the paper.
Network effects from social proof (reviews, UGC) can create winner-takes-most dynamics, advantaging destinations with stronger digital signals and creating visibility frictions for small/emerging destinations.
Theoretical argument drawing on platform/network effects literature and applied to tourism/social proof; paper cites social-proof constructs and suggests measurement via platform data.
Proprietary experimental datasets and curated metagenomic sequences become valuable intellectual assets that can differentiate commercial offerings.
Paper lists 'Data as an economic asset' and highlights the value of proprietary datasets and curated metagenomes; no market valuation data are included.
Faster, cheaper access to structural hypotheses can shorten drug and enzyme discovery cycles, raising R&D productivity and lowering marginal costs of early‑stage screening.
Paper argues this as an implication under 'Productivity and R&D acceleration'; it is presented as an economic consequence rather than demonstrated with empirical cost‑or time‑saving data in the text.
Practical applications are already emerging, including accelerating target structure availability for small‑molecule and biologics design, guiding enzyme redesign, and interpreting disease mutations.
Paper lists these application areas as emerging uses of AI‑predicted structures; evidence is presented as examples and implications rather than empirical case studies within the text.
Template‑and‑MSA informed architectures (e.g., RoseTTAFold and AlphaFold family) deliver near‑experimental accuracy for many proteins.
Paper names these architectures and links their inputs (MSAs, templates) to high accuracy against experimental structures (PDB); specific evaluation datasets, protein counts, or error metrics are not enumerated in the text.
Modern AI systems (e.g., AlphaFold variants, RoseTTAFold, single‑sequence models like ESMFold) can approach or reach near‑experimental accuracy while greatly increasing speed and scalability.
Paper cites specific models (AlphaFold family, RoseTTAFold, ESMFold) and describes benchmarking against structural ground truth (PDB / curated experimental structures) and large‑scale pretraining; exact benchmark values or sample sizes are not specified in the text.
New economic metrics are needed for VR (value of behavioral data streams, cost per reduction in harm, ROI on security investments, welfare metrics capturing trust and adoption).
Authors' recommendations based on identified gaps in the literature and the comparative review of 31 studies; proposed as agenda items rather than empirically developed metrics.
VR generates high‑value behavioral and biometric datasets for AI personalization, training, and analytics; firms that extract this data can gain competitive advantages, creating incentives to centralize collection unless counteracted by policy or market forces.
Economic implications inferred by the authors from the literature synthesis and standard industrial‑organization logic; not supported by original empirical market data in the paper.
There is a need for regulatory standards, industry best practices, and ethics‑by‑design approaches; interoperable policy frameworks are recommended to govern VR security and privacy.
Policy and governance recommendations synthesized from multiple reviewed studies and the authors' integration; presented as prescriptive guidance rather than empirically tested interventions.
An effective defense mix for VR combines technical controls (secure boot, attestation, encrypted communications), AI tools for anomaly detection and policy enforcement, and human‑centered design (transparency, consent, usable controls).
Cross‑study synthesis showing these categories recur as recommended controls in the 31 reviewed papers; authors propose combining them in TVR‑Sec. No deployment or performance metrics provided.
Socio‑Behavioral Safety measures (moderation, design constraints, psycho‑social safeguards) are necessary to prevent harassment, persuasion, addictive interfaces, and other psychological harms in shared virtual spaces.
Qualitative synthesis of social‑behavioral harms and proposed mitigations reported across the literature review (31 studies); comparative evaluation of socio‑technical controls.
User Privacy in VR requires managing highly sensitive behavioral and biometric traces with privacy‑preserving ML approaches (e.g., federated learning, differential privacy), consent mechanisms, and data minimization.
Repeated recommendations across the reviewed studies; authors synthesized privacy‑preserving technical approaches and governance mechanisms from the 31‑study corpus. No primary experiments demonstrating efficacy provided.
System Integrity defenses should cover hardware, firmware, sensors, and networks to protect against spoofing, device tampering, malware, and supply‑chain attacks.
Aggregated technical recommendations from the literature corpus (31 studies) and the authors' mapping of integrity threats to controls (secure boot, attestation, encrypted communications). No empirical testing of these controls in the paper.
The Three‑Layer VR Security Framework (TVR‑Sec) integrates System Integrity, User Privacy, and Socio‑Behavioral Safety into an adaptive, multidimensional defense architecture for VR systems.
Conceptual synthesis developed by the authors from a comparative literature review of 31 peer‑reviewed studies (2023–2025); framework created by mapping identified vulnerabilities to technical, AI, and human‑centered controls. No empirical validation or deployment testing reported.
Policy and platform design choices (e.g., provenance metadata, detection/disclosure of AI-generated content, monetization rule alignment) can reinforce or mitigate harms from GenAI-driven creator economies.
Policy recommendations and implications drawn from the qualitative findings across the 377-video sample and normative reasoning; not empirically tested.
Participants systematically over-bid for privacy-disclosure labels: they were willing to pay more for a privacy-disclosure label than its objective value.
In the experiment (N = 610) participants submitted willingness-to-pay bids for privacy-disclosure labels; observed bids exceeded the objective/reference value, indicating overpayment for transparency information.
Policy interventions that raise the reinstatement rate — for example, compensation/transfers to translate AI gains into broad-based purchasing power, faster/stronger fiscal support or automatic stabilizers — can prevent the explosive feedback and stabilize demand.
Model experiments and sensitivity analysis showing that increasing the reinstatement elasticity or direct transfers moves the system from explosive to convergent parameter regions in the calibrated phase-space.
FutureBoosting generalizes across multiple real-world electricity markets and forecast horizons.
Empirical results reported across 'multiple real-world electricity markets' and several forecasting horizons to capture diverse volatility and regime behavior (details on exact markets/horizons are reported in the experiments section of the paper).