Evidence (2066 claims)
Adoption
8570 claims
Productivity
7631 claims
Governance
6869 claims
Human-AI Collaboration
6491 claims
Org Design
4175 claims
Innovation
4114 claims
Labor Markets
3566 claims
Skills & Training
2966 claims
Inequality
2066 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 758 | 199 | 100 | 900 | 2007 |
| Governance & Regulation | 826 | 400 | 191 | 122 | 1563 |
| Organizational Efficiency | 777 | 193 | 124 | 84 | 1189 |
| Technology Adoption Rate | 635 | 233 | 124 | 97 | 1098 |
| Research Productivity | 422 | 128 | 57 | 336 | 954 |
| Output Quality | 476 | 179 | 59 | 47 | 761 |
| Decision Quality | 328 | 177 | 81 | 47 | 640 |
| Firm Productivity | 435 | 57 | 88 | 20 | 606 |
| AI Safety & Ethics | 218 | 277 | 65 | 33 | 599 |
| Market Structure | 180 | 170 | 123 | 24 | 502 |
| Task Allocation | 213 | 64 | 72 | 33 | 387 |
| Skill Acquisition | 170 | 61 | 61 | 17 | 309 |
| Innovation Output | 203 | 27 | 43 | 18 | 292 |
| Employment Level | 105 | 54 | 107 | 13 | 281 |
| Fiscal & Macroeconomic | 131 | 69 | 43 | 26 | 276 |
| Consumer Welfare | 117 | 63 | 42 | 11 | 233 |
| Firm Revenue | 153 | 48 | 26 | 3 | 230 |
| Task Completion Time | 173 | 31 | 8 | 12 | 225 |
| Inequality Measures | 44 | 122 | 49 | 6 | 221 |
| Worker Satisfaction | 89 | 65 | 22 | 12 | 188 |
| Error Rate | 69 | 92 | 10 | 2 | 173 |
| Regulatory Compliance | 77 | 69 | 14 | 5 | 165 |
| Automation Exposure | 56 | 56 | 26 | 13 | 154 |
| Training Effectiveness | 94 | 21 | 13 | 19 | 149 |
| Wages & Compensation | 77 | 36 | 25 | 6 | 144 |
| Team Performance | 86 | 17 | 27 | 10 | 141 |
| Developer Productivity | 95 | 17 | 14 | 6 | 133 |
| Job Displacement | 12 | 80 | 20 | 1 | 113 |
| Hiring & Recruitment | 52 | 7 | 8 | 3 | 70 |
| Creative Output | 31 | 18 | 8 | 3 | 61 |
| Skill Obsolescence | 5 | 46 | 6 | 1 | 58 |
| Social Protection | 27 | 16 | 8 | 2 | 53 |
| Labor Share of Income | 17 | 19 | 17 | — | 53 |
| Worker Turnover | 11 | 12 | — | 3 | 26 |
| Industry | — | — | — | 1 | 1 |
Inequality
Remove filter
This article develops the AI Infrastructure Triad as a conceptual framework for analyzing three competing priorities in regional AI infrastructure governance: Progress, Sustainability, and Equity.
Theoretical/conceptual development presented in the paper; synthesis of prior work on economic, physical, and moral limits of AI development.
Research on automation should be reoriented away from a primary focus on job loss toward understanding the organizational and technological transformations produced by digital work.
Normative and methodological recommendation derived from the paper's critical review of literature and the mappings of production/work networks; argued on conceptual and interpretive grounds rather than new empirical estimation.
The global HR technology market is expected to expand from USD 43.7 billion in 2025 to over USD 81 billion by 2032.
Forecast figure stated in paper (likely sourced from a market research / industry report, not specified in the excerpt).
Artificial Intelligence (AI) is increasingly marketed as a neutral arbiter capable of eliminating unconscious bias from human resource processes.
Statement in paper (assertion about industry marketing and positioning); no empirical data or citation provided in the excerpt.
Community and Indigenous approaches offer alternative models of authority over AI infrastructure rooted in stewardship rather than extraction, although these approaches are constrained.
Normative argument and engagement with community/Indigenous scholarship and examples; presented as an alternative model in the paper (qualitative).
Managerially, firms should pair GenAI access with short AIC micro-training and simple standard operating procedures (SOPs) to capture value consistently and avoid uneven adoption outcomes.
Authors' managerial recommendation drawn from experimental findings that AIC predicts gains and that scaffolding reduces variance; recommendation is an interpretation/synthesis rather than a directly tested organizational field intervention.
A scaffolding intervention (conceptual maps) reduced outcome variance, indicating that standardized workflows can mitigate inequality in AI-mediated performance.
Experimental inclusion of a scaffolding intervention (conceptual maps) and reported reduction in variance of outcomes among participants receiving scaffolding in conjunction with GenAI access.
Improvements were not predicted by GPA or prior knowledge, but were predicted by AI Interaction Competence (AIC) — the ability to elicit, filter, and verify model outputs.
Regression/subgroup analyses reported in the experiment linking improvements in task performance to measured predictors (GPA, prior knowledge, AIC); authors report null association for GPA/prior knowledge and positive association for AIC.
On average, GenAI access significantly increased task performance.
Reported randomized controlled experiment comparing task performance between LLM-assisted group and traditional-resources group; authors state the average increase was statistically significant.
We demonstrate its extraterritorial scope for gaining access to elements such as employment contracts and NDAs that have never been provided to the workers concerned.
Reported legal/empirical demonstration in paper: GDPR requests resulting in access to employment contracts and nondisclosure agreements (NDAs) that workers had not previously received. (Exact number of successful requests not stated in the excerpt.)
We audit the working conditions of content moderators in Kenya and Nigeria employed by business process outsourcing (BPO) companies by using the European General Data Protection Regulation (GDPR).
Method reported in paper: use of GDPR data-subject access / information requests to BPOs and platforms to obtain employment-related documents for content moderators in Kenya and Nigeria. (Sample size / number of requests not stated in the excerpt.)
The framework contributes to improving understanding of enterprise coordination and governance under constrained legal conditions and offers a basis for future analytical and empirical research.
Author-stated contribution of the paper based on the developed theoretical framework; positioned as foundation for future work.
The analysis identifies theoretical conditions under which such governance may support verifiable integrity, adaptive compliance, and access to formal markets.
Theoretical conditions derived from the review and theory synthesis (no empirical testing reported in this paper).
The study develops a theory-based framework explaining how RegTech-supported governance may, under specified conditions, enable sanctions-safe enterprise ecosystems during post-conflict reconstruction.
Primary contribution of the paper: theory synthesis built from integrative review of five literature streams (RegTech, sanctions compliance, institutional voids, supply-chain governance, algorithmic accountability).
Post-conflict reconstruction relies heavily on private enterprises to bring back employment, rebuild supply networks, and reconnect damaged economies.
Statement grounded in literature cited in the review (paper positions this as a general premise from post-conflict reconstruction literature); no primary data reported.
These results challenge the presumed universality of the fairness-accuracy tradeoff and demonstrate that well-designed modeling improvements can advance both fairness and accuracy in large-scale public sector systems.
Synthesis of the three complementary analyses (observational county-level correlations, simulation experiments with added property features, and simulations incorporating Census data) performed on the 26 million-sale dataset covering ~95% of U.S. counties.
Incorporating publicly available Census data into assessment models - a feasible reform in most counties - would significantly improve both accuracy and fairness relative to status quo assessments.
Simulated reforms adding publicly available Census covariates to assessment models and comparing resulting accuracy and fairness metrics to status-quo assessments across the dataset covering 26 million sales/95% of counties.
When accuracy improves in the simulated assessment models, fairness almost always improves as well.
Analysis of simulated model outcomes showing joint changes in accuracy and fairness metrics across many simulated configurations and counties; reported near-universal co-improvement when accuracy rises.
In simulated assessment models, adding property features improves accuracy in most cases.
Simulation experiments using alternative assessment models that include additional property-level features; comparisons between baseline and feature-augmented simulated models across many counties/cases.
Assessment accuracy and fairness - measured using domain-relevant metrics - are strongly correlated across counties under status quo practices.
Observational analysis of status-quo assessment outcomes using a dataset of 26 million property sales spanning ~95% of U.S. counties; county-level correlation analysis between domain-relevant accuracy metrics and fairness metrics.
Policy should prioritize employment‑centered digital strategies that are spatially differentiated and institutionally grounded to mitigate negative labor and development effects.
Normative policy recommendation arising from the paper's theoretical framework and regional field observations (policy prescription; not an empirically estimated intervention in the paper).
The paper proposes a reconstructed labour law framework based on economic dependency rather than traditional employment classification, including recognition of dependent contractor status, platform liability for worker welfare, algorithmic transparency, social security obligations, and specialised grievance mechanisms.
Normative legal/policy proposal articulated by the author(s) based on theoretical argument and the comparative analysis of existing regulatory gaps; prescriptive recommendation rather than empirically tested intervention.
Policy conclusion: while palliative care is an ethical imperative, its expansion must be decoupled from the oncological paradigm and matched with state-funded long-term care to protect against clinical decline and financial shocks.
Normative recommendation based on the empirical distributional findings (average protective effects but harmful tails for vulnerable groups) and cross-national differences reported in the analysis.
We introduce a Synthetic Data Generation framework using Tabular Denoising Diffusion Probabilistic Models within a Two-Learner architecture to synthesize high-fidelity digital twins from pan-European SHARE data (2016-2021).
Methodological contribution described in the paper; implementation details include use of diffusion-based tabular generative models and a Two-Learner architecture applied to SHARE microdata from 2016–2021.
On average, palliative care (PC) acts as a 'double shield', truncating out-of-pocket expenditures (financial toxicity) and informal caregiving shadow values (time poverty).
Analysis of pan-European SHARE data (2016-2021) using a Synthetic Data Generation framework (Tabular Denoising Diffusion Probabilistic Models within a Two-Learner architecture) to create digital twins and estimate treatment effects.
The study highlights the importance of reskilling and education reforms to ensure inclusive labor market outcomes in the era of AI-driven transformation.
Authors' policy recommendation based on their empirical findings from the survey (n=320) and SEM analysis; presented as a conclusion/recommendation rather than a quantified empirical result.
The model explained 49% of variance in wage dynamics (R^2 = 0.49).
SEM model statistics reported for the survey-based model (n=320); R-squared for wage dynamics = 49%.
The model explained 45% of variance in skill transformation (R^2 = 0.45).
SEM model statistics reported for the survey-based model (n=320); R-squared for skill transformation = 45%.
The model explained 52% of variance in employment patterns (R^2 = 0.52).
SEM model fit/variance-explained statistics reported for the survey-based model (n=320); R-squared for employment patterns = 52%.
Mediation analysis confirmed that skill transformation plays a significant mediating role linking AI adoption with wage distribution/outcomes.
Mediation analysis within the SEM framework applied to the survey data (n=320); authors report a significant mediation effect (no numeric indirect effect reported in the summary).
Mediation analysis confirmed that skill transformation plays a significant mediating role linking AI adoption with employment outcomes.
Mediation analysis within the SEM framework applied to the survey data (n=320); authors report a significant mediation effect (no numeric indirect effect reported in the summary).
Skill transformation significantly affected wage dynamics (β = 0.55, p < 0.001).
Structural equation modeling (SEM) on the same sample (n=320); reported standardized path coefficient β = 0.55 with p < 0.001.
Skill transformation significantly affected employment patterns (β = 0.58, p < 0.001).
Structural equation modeling (SEM) mediation/causal-path analysis on the survey (n=320); reported standardized path coefficient β = 0.58 with p < 0.001.
AI adoption significantly influenced wage dynamics (β = 0.61, p < 0.001).
Structural equation modeling (SEM) on the same survey sample (n=320); reported standardized path coefficient β = 0.61 with p < 0.001.
AI adoption significantly influenced skill transformation (β = 0.67, p < 0.001).
Structural equation modeling (SEM) on the same survey sample (n=320); reported standardized path coefficient β = 0.67 with p < 0.001.
AI adoption significantly influenced employment patterns (β = 0.63, p < 0.001).
Structural equation modeling (SEM) on primary survey data from n=320 employees across IT, banking, manufacturing, education, and service sectors; reported standardized path coefficient β = 0.63 with p < 0.001.
The model identifies simple measures/conditions that characterize when productivity paradoxes and skill polarization arise.
Theoretical derivations and analytical characterizations within the model yielding threshold conditions and measures parameterizing when paradoxical outcomes occur (model-based; no empirical validation).
This budget-split approach is responsive to the needs of real-world, resource-constrained advertisers committed to equitable distribution of public service outreach via online advertising.
Authors' normative/qualitative conclusion based on the implemented intervention and its practical suitability for government advertisers; no empirical quantification provided in excerpt.
The budget split intervention is a valuable approach to addressing ad delivery skew without excluding unknown users.
Authors' empirical finding from the collaboration/intervention (paper reports results from implemented intervention; specific metrics, sample size, and quantitative results are not provided in the excerpt).
In the absence of platform-provided solutions to skewed ad delivery, advertisers can counteract skew by targeting demographic groups directly.
Descriptive claim about common advertiser strategies; motivated by platform capability gaps (no experimental/sample details in excerpt).
There is a 15%–22% wage premium for workers demonstrating AI-augmentation capabilities.
Reported range across synthesized empirical studies documenting wage differences associated with demonstrated AI-augmentation capabilities.
The study draws policy implications for EU Cohesion programming and Sustainable Development Goals 4, 8, 9, 10, and 17.
Paper explicitly states policy implications and links to specific SDGs in its conclusions.
External technology partnerships, targeted education, and economic incentives operate as enablers [of AI adoption], all mediated by social and human capital availability.
Thematic analysis of interview data identifying these factors as enabling AI adoption, with mediation by social/human capital.
The results generalize existing optimality theorems for fairness-constrained classification and extend them to generalized fairness metrics and partial fairness regimes.
Mathematical generalization and extension of prior theorems to a broader class of fairness metrics and to settings with partial (not full) fairness constraints; proofs provided in the paper.
This result complements existing optimality theorems from the literature which, for specific fairness constraints, posit lower-bound threshold rules only.
Comparative theoretical discussion and extension of prior optimality results (literature comparison plus proofs showing how their characterization extends prior lower-bound-only threshold results).
The Pareto frontier consists of deterministic, group-specific threshold rules applied to individuals' success probability.
Theoretical analysis framing decision making as a multi-objective optimization problem (decision-maker utility vs. group fairness) and deriving the set of Pareto-optimal decision rules for arbitrary utility functions, arbitrary population distributions, and a wide range of group fairness metrics (mathematical proofs/derivations).
A majority seems optimistic about [AI's] overall impact.
Paper reports a majority-level positive attitude in surveys about AI's overall impact (no survey details or sample sizes provided in the excerpt).
Hallucinated references disproportionately assign credit to already prominent and male scholars, suggesting LLM-generated errors may reinforce existing inequities in scientific recognition.
Analysis linking hallucinated citations to characteristics of the (intended or assigned) cited authors, including measures of prominence and inferred gender, showing over-representation of prominent and male scholars among hallucinated attributions.
Hallucinated references are especially pronounced among small and early-career author teams.
Analysis of hallucination prevalence by author-team characteristics (team size and author career stage) within the audited dataset.
Hallucinated references are especially pronounced in manuscripts with linguistic signatures of AI-assisted writing.
Classification of manuscripts by linguistic features (signatures) indicative of AI-assistance and comparison of hallucination prevalence between groups.