Evidence (14922 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
9047 claims
Filter claims →
Productivity
8066 claims
Filter claims →
Governance
7278 claims
Filter claims →
Human-AI Collaboration
6912 claims
Filter claims →
Org Design
4439 claims
Filter claims →
Innovation
4359 claims
Filter claims →
Labor Markets
3652 claims
Filter claims →
Skills & Training
3018 claims
Filter claims →
Inequality
2160 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 | 795 | 210 | 105 | 955 | 2131 |
| Governance & Regulation | 886 | 414 | 197 | 126 | 1654 |
| Organizational Efficiency | 826 | 204 | 129 | 87 | 1257 |
| Technology Adoption Rate | 681 | 259 | 128 | 110 | 1189 |
| Research Productivity | 464 | 138 | 65 | 349 | 1028 |
| Output Quality | 503 | 196 | 61 | 53 | 813 |
| Decision Quality | 351 | 180 | 84 | 51 | 673 |
| AI Safety & Ethics | 238 | 288 | 71 | 34 | 637 |
| Firm Productivity | 455 | 58 | 92 | 20 | 631 |
| Market Structure | 186 | 172 | 123 | 25 | 511 |
| Task Allocation | 222 | 70 | 76 | 34 | 407 |
| Innovation Output | 238 | 28 | 48 | 18 | 334 |
| Skill Acquisition | 177 | 62 | 62 | 17 | 318 |
| Employment Level | 107 | 57 | 108 | 13 | 287 |
| Fiscal & Macroeconomic | 135 | 72 | 44 | 26 | 284 |
| Firm Revenue | 172 | 50 | 28 | 5 | 256 |
| Consumer Welfare | 121 | 68 | 45 | 12 | 246 |
| Task Completion Time | 183 | 33 | 10 | 13 | 240 |
| Inequality Measures | 45 | 126 | 50 | 6 | 227 |
| Worker Satisfaction | 95 | 74 | 23 | 12 | 204 |
| Error Rate | 77 | 98 | 11 | 4 | 190 |
| Regulatory Compliance | 84 | 73 | 17 | 7 | 181 |
| Automation Exposure | 61 | 61 | 27 | 14 | 166 |
| Training Effectiveness | 98 | 21 | 14 | 19 | 154 |
| Wages & Compensation | 78 | 37 | 25 | 6 | 146 |
| Developer Productivity | 105 | 18 | 14 | 6 | 144 |
| Team Performance | 87 | 17 | 28 | 10 | 143 |
| Job Displacement | 12 | 83 | 23 | 1 | 119 |
| Hiring & Recruitment | 53 | 8 | 8 | 3 | 72 |
| Social Protection | 39 | 17 | 8 | 2 | 66 |
| Creative Output | 32 | 20 | 8 | 3 | 64 |
| Skill Obsolescence | 5 | 50 | 6 | 1 | 62 |
| Labor Share of Income | 17 | 20 | 17 | — | 54 |
| Worker Turnover | 15 | 15 | — | 3 | 33 |
| Industry | — | — | — | 1 | 1 |
Beyond technological efficiency, AI carries the potential to strengthen societal welfare.
Normative assertion made in the paper (argumentative/literature-based); no specific empirical study, metrics, or sample size provided in the abstract.
Organizational adoption follows a diffusion-like process: Enthusiasts push ahead with tools, creating organizational success that converts Pragmatists.
Aggregated survey observations indicating teams or organizations with higher representation of 'Enthusiasts' report more tool uptake and subsequent increased adoption among 'Pragmatists'; based on self-reported organizational-level indicators from the 147-developer sample.
LLM-based chatbots may offer a means to provide better, faster help to nonprofit caseworkers assisting clients with complex program eligibility.
Motivating claim in introduction/abstract: potential for LLM-based chatbots to assist caseworkers; supported in the paper by experimental findings showing accuracy improvements with higher-quality chatbots, but not a direct field-deployment test of speed or real client outcomes.
Addressing these inequities through social protection may be particularly promising to achieve longer-term poverty-reduction goals, increase productive efficiency, and promote a better, more sustainable future.
Conditional/forward-looking claim made by the authors in the introduction; presented as a plausible policy pathway rather than supported here by specific empirical results (the chapter will review relevant evidence).
At a model size of 200M parameters, environment overhead is below 4% of training time.
Measured training time breakdowns at 200M-parameter models showing environment (simulation) overhead contribution under 4%. (Implied across their translated environments during benchmarking/training runs.)
Machine learning has potential to advance occupational health research if its capabilities are fully leveraged through interdisciplinary work.
Implied conclusion from the review's discussion and recommendation (the paper frames ML as having 'potential' if combined with interdisciplinary efforts; direct empirical evidence of realized advancement not provided in the excerpt).
Interdisciplinary collaboration is necessary to fully leverage the potential of machine learning in advancing occupational health research.
Conclusion/recommendation drawn by the paper's authors based on their review of the literature (stated as a need in the paper; empirical demonstration of this necessity is not provided in the excerpt).
Intelligent centralized orchestration fundamentally improves multimodal AI deployment economics.
Authors generalize from the reported empirical results (reductions in time-to-answer, conversational rework, and cost on their 2,847-query evaluation) to claim broader economic benefits of centralized orchestration.
Critical thinking development and ethical reasoning cultivation retain 70-75% human centrality.
Authors provide a numerical estimate (70-75% human centrality) in their functional analysis; the paper does not report empirical methods or sample evidence for this figure.
Mentorship and social development remain largely human-dependent with only 25-30% substitutability by AI.
Paper's estimated substitutability range (25-30%) for mentorship and social development; the estimate is not accompanied by empirical data or described methodology.
Future research should track long-term adoption trends, evaluate policy incentives, and integrate sustainability metrics to inform climate-resilient and inclusive agricultural innovation.
Paper's stated research agenda and recommendations for follow-up studies (qualitative, prospective).
Peer-driven digitalization matters not only for firm-level resilience but also for long-term sustainable competitiveness in manufacturing ecosystems.
Synthesis and implication drawn from empirical results (peer effects, mediators, and heterogeneity) using Chinese manufacturing A-share firm data from 2013–2022.
The adoption of AI technologies offers a scalable, resilient strategy for modernizing water management and promoting agricultural sustainability in Iraq.
Authors' conclusion based on single-site field experiments, economic and sustainability analyses, and reported robustness in sensitivity analyses; scalability claim is inferential and extends beyond the experimental site.
The results highlight the promise of incorporating public input into AI governance.
Authors' conclusion based on experimental findings that informational exposure can change public attitudes about AI in public decision contexts even when direct experience does not.
Future improvements in navigation and AI detection are expected to further enhance efficiency and adaptability of the weeder.
Authors' prospective recommendation based on current system performance and identified limitations; forward-looking statement rather than an empirical result.
The future of work must be human-centric, balancing technological efficiency with dignity, inclusion, and meaningful employment.
Normative conclusion/recommendation drawn by the authors from their conceptual and analytical discussion; not supported by original empirical testing within this paper.
Information Systems (IS) research is critical for achieving joint optimization of technical capabilities and social systems in the context of GenAI.
Authors' argumentative positioning based on the socio-technical interpretation of the review; proposed role for IS scholarship rather than empirical test within the review.
The presented framework contributes to the responsible use of AI, productivity, and long-term economic competitiveness in the United States.
Forward-looking claim rooted in conceptual reasoning and literature synthesis; no longitudinal data, economic modeling, or empirical evidence is provided to demonstrate the claimed macroeconomic effects.
A proactive approach (ensuring AI literacy and integrating best practices) will enable the workforce to effectively leverage AI technologies and remain resilient in an increasingly dynamic economic environment.
Projected outcome and recommendation in the paper's conclusion; presented as expected benefit rather than demonstrated result in the excerpt.
Career optimism can be positioned as an indicator of workforce sustainability and a strategic lever for innovation, with implications for organizations, educators, and policymakers aiming to cultivate resilient, future-ready labor markets.
Interpretation and recommendations in the paper's discussion section, drawing on the survey findings (associations between career optimism and organizational/regional factors) to argue for practical applications.
Deterministic verifiers and benchmarks like SkillsBench are important for certification and procurement decisions because they enable verifiable, repeatable gains.
Normative implication in the paper based on the use of deterministic verifiers to measure Skill impact reproducibly; this is an interpretive claim about downstream decision-making rather than an experiment-derived metric.
Focused, modular Skill design favors modular pricing and bundling strategies (i.e., narrow high-impact Skills premium; broad libraries lower margin).
Policy/market implication derived from the experimental finding that focused 2–3-module Skills outperform comprehensive documentation; the pricing/bundling claim is an economic inference, not empirically tested in the paper.
Because curated Skills yield large average gains, human curation of high-quality procedural knowledge has economic value and could be a high-return activity.
Paper's economic implication drawn from the empirical +16.2 pp average pass-rate improvement for curated Skills. This is an interpretation/inference rather than a direct empirical economic measurement.
Policy tools such as bans on sale of certain sensitive data, fiduciary duties for data holders, privacy-by-default, and collective data governance (data trusts, regulated commons) are appropriate levers to limit harms from data commodification.
Prescriptive policy argument based on normative analysis and literature on governance alternatives; recommendations are not evaluated using empirical policy impact studies within the paper.
Policy-relevant implication (extrapolated): diffusion of AI tools among small firms will likely follow social-network channels and be shaped by peer benchmarking, so aggregate incentives may underperform unless they leverage local networks and trusted intermediaries.
Inference and policy implication drawn from main empirical findings on the primacy of social networks and peer effects for entrepreneurial behavior; not directly measured in the dataset for AI-specific adoption.
TVET-aligned training with portable, employer‑recognised credentials can change how employers value pre‑departure training—potentially raising match quality, wage outcomes, and mobility options.
Theoretical/signalling argument supported by policy instruments review and recommended employer-focused tests (surveys, hiring experiments); not empirically demonstrated in this paper.
Earlier, decentralised training with digital support could reduce search frictions and brokerage rents by improving migrants’ information and bargaining capacity (economic role).
Economic reasoning and conceptual linkage between information provision and transaction costs; suggested empirical strategies (RCTs/quasi-experiments) to test the claim but no causal estimates reported.
Proposition 2: TVET alignment and portable skills recognition (functional, employer‑usable verification such as micro‑credentials) let training convert into labour‑market value and mobility options.
Policy-analytic argument supported by review of recognition/QA instruments and transferability concepts; paper recommends employer surveys and hiring experiments to test this but provides no causal evidence.
Proposition 1: Earlier, decentralised access to training reduces information asymmetry and dependence on intermediaries.
Presented as a testable proposition derived from corridor process mapping and conceptual analysis; recommended for randomized or quasi-experimental evaluation but not empirically tested in this paper.
Redesigning pre-departure training along four axes—standards, timing, delivery architecture, and recognition/portability—can reduce information asymmetries, lower dependence on brokers, and better connect migration to labour‑market value without waiting for slower permit/enforcement reforms.
Argument derived from conceptual reframing and corridor process mapping; supported by desk review and governance gap analysis. Presented as a policy proposition rather than empirically tested causal claim.
China exhibits strong long-run integration between core AI and AI-enhanced robotics and a significant contribution from universities and the public sector to patenting.
Country-level decomposition showing (a) a stronger statistical long-run relationship between Chinese core AI and AI-enhanced robotics patent series and (b) actor-type decomposition of Chinese patent filings indicating relatively high shares from universities/public-sector actors (patents 1980–2019). Exact counts/shares not provided in the summary.
The system facilitates scenario and counterfactual analysis (e.g., education subsidies, AI taxation, adoption incentives) to stress-test policy options and firm-level responses under alternative diffusion scenarios.
Modeling proposal: task-based microsimulation and scenario ensembles are described as part of the architecture; no example counterfactual simulations or sample results are included.
The proposed phased implementation (pilots, holdouts, continuous validation, transparency) can be operationally integrated into BLS projection workflows.
Practical rollout plan described (phased pilots, backtesting, operational integration); this is a suggested implementation pathway rather than demonstrated integration. No implementation sample or timeline is provided.
Policymakers should combine competition policy, data governance, retraining/redistribution measures, and targeted R&D/green-AI incentives to manage the transition and preserve broad-based demand.
Normative policy recommendation derived from the integrated theoretical framework and literature synthesis; not empirically validated in the paper.
Economically, there will be demand for 'temporal-quality' products: neurotech and AI services that explicitly measure, preserve, or enhance experienced temporality (presence, flow, meaning), representing a distinct market segment.
Speculative market implication derived from conceptual argument and literature on consumer preferences; no market data or empirical demand studies provided.
Recommended priorities include funding longer, practice‑embedded programs, developing standardized competency frameworks and validated assessments, and conducting studies that link training to organizational and patient outcomes (to enable level‑4 evidence and economic evaluation).
Authors' practical and policy recommendations based on synthesis of findings (limited depth/duration of current programs and lack of level‑4 outcomes) described in the paper.
Interpretive claim: AI interventions (upskilling and AI-guided workflows) raise worker confidence and job satisfaction and help tailor stress-management approaches, which can support retention under stress.
Authors' interpretive summary (not tied to a specific reported coefficient); described as a mechanism for the observed AI moderation on retention. Instrument/scale details and direct measurement of confidence/job satisfaction not provided in the summary.
Respondents recommend co-designing policies and curricula with educators and students, prioritizing hands-on low-cost training (open-source tools, cloud credits, shared labs), and investing in pooled infrastructure with targeted support for under-resourced regions.
Recurring recommendations identified through thematic coding of open-ended survey responses and synthesis of respondent suggestions; supportive quantitative items indicating preferences for specific interventions.
To establish causal links between price, perceived value, and outcomes, researchers should use field experiments, A/B tests, instrumental variables, and natural experiments.
Methodological recommendations in the paper's implications section, grounded in authors' assessment of current methodological gaps.
AI economics research should build hybrid behavioral–machine learning models that predict perceived value at scale and integrate them into pricing optimization frameworks.
Implications and research agenda provided by the authors based on gaps identified in the SLR; recommended modeling approach rather than empirical finding.
Future research should incorporate ethics, fairness, and transparency into pricing algorithms and leverage predictive technologies to estimate and operationalize perceived value in real time.
Authors' explicit future-research recommendations derived from gaps identified in the SLR.
Organizational capabilities (data, analytics, governance, cross-functional alignment) are critical enablers of successful digital VBP.
Repeated identification of organizational capability factors across the 30 reviewed studies and synthesis into a thematic cluster by the authors.
Continuous CPD records enable predictive models for upskilling needs; AI can personalize training pathways and recommend CPD courses that maximize employability or wage growth.
Projected application described in the AI-economics implications; not empirically tested in the paper.
Automated compliance and auditable dashboards can lower transaction costs and improve matching efficiency between employers and certified technicians/engineers.
Conceptual argument drawing on transaction-cost economics and system design; no measured changes in transaction costs or matching outcomes reported.
Standardized, machine-readable records enable credential portability and lower verification costs for employers and platforms.
Theoretical argument in the paper's implications section; no empirical evidence or cost-estimates provided.
Digitized, cloud-hosted credential records would create high-quality administrative datasets that AI can use to model career trajectories, estimate returns to credentials, and automate verification—reducing signalling frictions in labour markets.
Policy/AI-economics implications argued in the paper; forward-looking claim based on expected properties of machine-readable administrative data, not empirical demonstration.
Industrial automation (industrial robots) can be an effective component of green development strategies when paired with finance and policy instruments.
Inference drawn from core empirical results: (1) IR reduces IWE; (2) effects are stronger with greater financial depth and policy support; combined evidence suggests complementarity between automation, finance, and policy.
Regulators must balance innovation with consumer protection by mandating model auditability, fairness testing, and interoperable data standards to prevent systemic and algorithmic risks.
Policy recommendation derived from synthesis of algorithmic risk, model opacity, and fintech market dynamics; based on normative analysis and best‑practice proposals rather than empirical testing.
Observed higher short-term performance and the positive correlation with iterative engagement imply that GenAI can augment short-term academic productivity and that benefits depend partly on active, skillful user interaction (complementarity).
Synthesis in implications drawing on the experimental finding of higher scores for allowed-use groups and the positive correlation between number of edits and performance; this interpretive claim is inferential and not directly tested as a structural complementarity in the study.
The FutureBoosting hybridization approach can be generalized to other economic time-series forecasting tasks (e.g., macro indicators, commodity prices, demand forecasting).
Paper's implications and discussion section proposing generalization; conceptual argument rather than direct empirical evidence in non-electricity domains.