Evidence (8066 claims)
Adoption
5586 claims
Productivity
4857 claims
Governance
4381 claims
Human-AI Collaboration
3417 claims
Labor Markets
2685 claims
Innovation
2581 claims
Org Design
2499 claims
Skills & Training
2031 claims
Inequality
1382 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 417 | 113 | 67 | 480 | 1091 |
| Governance & Regulation | 419 | 202 | 124 | 64 | 823 |
| Research Productivity | 261 | 100 | 34 | 303 | 703 |
| Organizational Efficiency | 406 | 96 | 71 | 40 | 616 |
| Technology Adoption Rate | 323 | 128 | 74 | 38 | 568 |
| Firm Productivity | 307 | 38 | 70 | 12 | 432 |
| Output Quality | 260 | 71 | 27 | 29 | 387 |
| AI Safety & Ethics | 118 | 179 | 45 | 24 | 368 |
| Market Structure | 107 | 128 | 85 | 14 | 339 |
| Decision Quality | 177 | 75 | 37 | 19 | 312 |
| Fiscal & Macroeconomic | 89 | 58 | 33 | 22 | 209 |
| Employment Level | 74 | 34 | 78 | 9 | 197 |
| Skill Acquisition | 98 | 36 | 40 | 9 | 183 |
| Innovation Output | 121 | 12 | 24 | 13 | 171 |
| Firm Revenue | 98 | 35 | 24 | — | 157 |
| Consumer Welfare | 73 | 31 | 37 | 7 | 148 |
| Task Allocation | 87 | 16 | 34 | 7 | 144 |
| Inequality Measures | 25 | 76 | 32 | 5 | 138 |
| Regulatory Compliance | 54 | 61 | 13 | 3 | 131 |
| Task Completion Time | 89 | 7 | 4 | 3 | 103 |
| Error Rate | 44 | 51 | 6 | — | 101 |
| Training Effectiveness | 58 | 12 | 12 | 16 | 99 |
| Worker Satisfaction | 47 | 33 | 11 | 7 | 98 |
| Wages & Compensation | 54 | 15 | 20 | 5 | 94 |
| Team Performance | 47 | 12 | 15 | 7 | 82 |
| Automation Exposure | 27 | 26 | 10 | 6 | 72 |
| Job Displacement | 6 | 39 | 13 | — | 58 |
| Hiring & Recruitment | 40 | 4 | 6 | 3 | 53 |
| Developer Productivity | 34 | 4 | 3 | 1 | 42 |
| Social Protection | 22 | 11 | 6 | 2 | 41 |
| Creative Output | 16 | 7 | 5 | 1 | 29 |
| Labor Share of Income | 12 | 6 | 9 | — | 27 |
| Skill Obsolescence | 3 | 20 | 2 | — | 25 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
AI and automation pose significant challenges to employment stability, skill relevance, and human dignity.
Claim presented within the paper's conceptual and analytical discussion of AI's dual impacts; no empirical study, sample size, or quantitative measures provided in this paper.
Combined analysis using Fuzzy PROMETHEE II and DEMATEL identifies High Initial Investment and Supply Chain Integration as critical barriers and dominant causal drivers that influence other dependent barriers.
Findings come from the integrated PROMETHEE II ranking and DEMATEL causal-mapping analyses based on expert input and literature review; detailed sample size and numerical results not provided in the summary.
Information processing constraints hinder managers' ability to effectively integrate tax planning and core business strategies (i.e., processing constraints hinder effective tax planning).
The paper reports novel empirical evidence consistent with this theoretical claim based on observed associations and tests linking AI, information quality, capital management, and tax effectiveness in the 2010–2018 sample.
Jurisdictions that implemented employee classification requirements experienced an 18% reduction in platform labor supply.
Comparative policy analysis across jurisdictions within the 24-country dataset comparing platform labor supply before and after employee-classification reforms using administrative and platform transaction records.
Median gig-worker hourly pay ($14.20) is approximately 22% below comparable traditional employment wages.
Comparison of adjusted median hourly gig earnings (platform records) to comparable hourly wages in traditional employment from labor force and administrative wage data for the same populations across the 24 countries.
There are challenges to adopting AI in HRM within IT firms.
Identified through the literature review and the empirical study involving HR professionals; the summary notes challenges but does not enumerate or quantify them.
Performance expectancy is a negative factor related to the company's decision to adopt AI (attributed to initial implementation challenges reducing perceived ease of use).
PLS-SEM analysis of survey data from 207 firms; the paper reports a negative association between performance expectancy and AI Adoption and offers a rationale about 'reality check' and initial implementation difficulties.
Concerns about privacy risks, overreliance on technology, and decision fatigue continue to shape consumer trust and adoption of AI features.
Reported qualitative/quantitative findings from the questionnaire and analysis indicating these concerns emerged as factors affecting trust and adoption (specific measurement items and effect sizes not reported in the summary).
LLM explanations foster inappropriate reliance and trust on the data-extraction AI: participants were less likely to detect errors when provided with LLM explanations.
User study measuring error-detection rates and trust/reliance indicators across conditions (full text, passage retrieval, LLM explanations). The LLM-explanation condition showed lower error-detection and greater reliance/trust compared to other conditions.
Governance quality becomes negative and statistically significant at the 0.90 quantile (τ = 0.90), which the paper interprets as evidence of institutional rigidity in advanced financial systems.
MMQR results showing a negative, significant coefficient for governance quality at τ = 0.90; interpretation provided by the authors linking this sign to institutional rigidity.
AI use also poses risks, including systemic discrimination, privacy invasion, and commodification of talent.
Qualitative synthesis and documented instances in the reviewed literature (n=85) reporting discriminatory outcomes, privacy concerns, and labor commodification effects associated with algorithmic HR tools.
Qualitative synthesis reveals a 'gray zone' in labor relations and a 'black box' in algorithmic data processing, both exposing businesses to procedural injustice risks.
Thematic/qualitative synthesis of findings from the reviewed literature (n=85) highlighting issues of labor relations and algorithmic opacity leading to procedural fairness concerns.
Digital transformation raises challenges related to privacy, inequality, and regulatory scrutiny.
Identified as a key challenge in the paper; the abstract provides no details on how privacy concerns, inequality measures, or regulatory incidents were documented or quantified.
We lack frameworks for articulating how cultural outputs might be actively beneficial.
Authors' identification of a gap in evaluation theory and practice (conceptual analysis); no systematic literature review details provided in the excerpt.
Current AI evaluation practices show a critical asymmetry: while AI assessments rigorously measure both benefits and harms of intelligence, they focus almost exclusively on cultural harms.
Authors' review/ critique of existing evaluation frameworks and metrics (qualitative analysis in the paper); the excerpt does not list the reviewed studies or their number.
The field of AI is unprepared to measure or respond to how the proliferation of entertaining AI-generated content will impact society.
Authors' assessment of current evaluation practices and frameworks (qualitative analysis presented in the paper); no empirical metrics or sample sizes provided in the excerpt.
Interpreting the literature through a socio-technical lens reveals a persistent misalignment between GenAI's fast-evolving technical subsystem and the slower-adapting social subsystem.
Authors' conceptual interpretation of the reviewed studies (28 papers) using socio-technical theory to integrate technical and social themes from the literature.
Evidence strength is inversely correlated with intervention complexity.
Cross-domain synthesis reported in the paper that formalises an inverse evidence–complexity relationship based on the reviewed literature. The abstract does not quantify the correlation or list the domains/intervention types used to derive it.
Per-capita elderly care costs running 3–5 times those of working-age cohorts.
Cost comparisons reported in sources included in the 81-paper review. The abstract reports a 3–5x multiple but does not specify which cost categories, countries, or methodological adjustments were used.
Conventional policy instruments have failed to resolve pressures that include severe long-term care workforce shortfalls across leading ageing economies.
Synthesis of findings from the structured narrative review of 81 sources (2020–2025) indicating persistent workforce shortfalls. The abstract does not provide quantitative workforce shortfall magnitudes or country-specific data.
Demographic ageing is projected to reduce annual GDP growth by 0.3–1.2 percentage points by 2035.
Projection estimates referenced in the review literature (2020–2025). The abstract reports the 0.3–1.2 p.p. range but does not specify which models or studies generated these projections.
Ageing-related expenditure already absorbs up to 18% of GDP in the most affected economies.
Spending estimates drawn from the reviewed literature (2020–2025). The paper states 'up to 18% of GDP' for the most affected economies but does not list which economies or the original data sources in the abstract.
Advanced economies face a compounding demographic crisis: populations aged 65 and over will reach 30–40% in several nations by 2050.
Demographic projection claims cited in the paper's background literature (sources from the structured narrative review). No specific datasets or country-by-country breakdown provided in the abstract.
Current literature has primarily focused on automation-based views of decision support and lacks insight into systematic human–AI coordination aided by analytics.
Literature review and conceptual critique within the paper. No systematic mapping study or bibliometric counts reported.
Most organizations have difficulties converting algorithmic results into sustainable managerial decisions due to low levels of trust, lack of explanation, and poor integration between AI systems and human judgment.
Synthesis of existing literature presented in the conceptual paper (literature review). No empirical study or sample provided to quantify 'most organizations.'
AI adoption has augmented complexity, uncertainty in decision-making, and accountability stresses for managers.
Claim supported by conceptual argument and literature integration (qualitative synthesis). No empirical sample size or quantitative testing reported.
Traditional methods for assessing and developing employees' skills often fail to provide real-time feedback.
Statement supported by literature review cited by the authors; the abstract does not provide empirical comparisons, metrics, or sample sizes.
The findings constitute a cautionary case for the effectiveness of LLM use in strategic decision-making.
Authors' interpretation based on the experimental results: representational changes occurred with LLM use but did not translate into improved strategic foresight, combined with observed increases in overload and decreases in ownership.
LLM use reduces psychological ownership (additional analyses).
Reported follow-up/additional analyses from the experiment showing a statistically significant decrease in psychological ownership measures for participants using LLMs.
Existing research on AI-driven decision-making remains fragmented and often framed through substitution-oriented narratives that position AI as a replacement for human judgment.
Assessment based on the author's interdisciplinary literature synthesis (conceptual meta-analysis); descriptive evaluation of research framing rather than new empirical testing.
Skills mismatch and SME adoption constraints constitute a binding bottleneck for inclusive digital–green upgrading.
Synthesis of studies on skills, firm capabilities, and SME adoption of digital and green technologies (review-level evidence; no single dataset or sample size provided).
Absent complementary institutions and infrastructure, digitalization may increase electricity demand, widen inequality, and incentivize strategic disclosure (greenwashing).
Literature review drawing on empirical studies of energy consumption from digital systems, labor-market studies, and analyses of ESG disclosure practices (review-level synthesis; no single sample size reported).
The IT sector is currently witnessing significant workforce restructuring, including employee layoffs, necessitating a critical reassessment of existing competency mapping frameworks.
Asserted in the paper as a motivating observation; no specific layoffs data or statistics provided in the excerpt.
More experienced translators appear more likely to exit the market after ChatGPT’s launch than less experienced translators.
Heterogeneous (subgroup) analysis by experience level within the translation market reported in the paper; evidence presumably from DiD estimates of exit/participation rates across experience levels. (Exact sample sizes and exit definitions not provided in the abstract.)
Following ChatGPT’s launch, some online labor markets experienced displacement effects characterized by reduced work volume and earnings, exemplified by the translation & localization OLM.
Empirical analysis using a Difference-in-Differences (DiD) design on online labor market (OLM) data; the abstract identifies translation & localization OLM as an example. (Sample size and exact data window not specified in the abstract.)
The review identifies highly heterogeneous modeling approaches with limited convergence toward shared benchmark tasks.
Comparative assessment across the 42 studies indicating a wide variety of modeling choices and an absence of commonly adopted benchmark tasks for direct comparison.
The literature reveals constraints, including challenges in processing long financial documents, limited availability of labeled datasets, and strong geographic and linguistic concentration.
Synthesis of methodological limitations and practical constraints reported across the reviewed studies (issues repeatedly mentioned in the corpus of 42 studies).
Embedding-based representations and end-to-end deep learning architectures appear only sporadically.
Review observations that only a small subset of the 42 studies used embedding representations or end-to-end deep learning models, i.e., these approaches are uncommon in the sample.
Less attention has been given to how sentiment-based textual features obtained from corporate reports are integrated into machine learning pipelines to predict firms' financial outcomes.
Synthesis from the systematic review of 42 studies indicating relatively few studies use corporate report–derived sentiment or explicitly address integration of such textual features into ML pipelines for firm-level financial predictions.
The AI productivity paradox reflects organizational constraints rather than technological failure.
Synthesis of the theoretical productivity funnel and empirical findings from firm-level data across Serbia, Croatia, Czechia, and Romania indicating conditional (not universal) productivity effects of AI.
Measurable productivity gains remain modest for firms lacking standardized processes and management systems.
Empirical comparisons within the firm-level dataset showing smaller productivity gains among firms characterized as lacking standardized processes/management systems (organizational readiness measures).
Within this framework, we identify a complementarity trap: firms lacking organizational readiness become stuck in the funnel, unable to convert AI diffusion into productivity gains.
Theoretical argument supplemented by empirical analysis using firm-level data from a subset of Central and Eastern European economies and AI diffusion indicators (countries named: Serbia, Croatia, Czechia, Romania).
The system forces many children to age out at 21, creating deportation risks for those who are American in every meaningful sense except paperwork.
Policy consequence of long backlogs: derivative status rules cause dependents to 'age out' at 21; deportation risk implication is a legal/administrative outcome. The excerpt does not quantify the number affected or present a dataset.
The backlog traps H-4 dependent spouses, over 90% of whom hold bachelor's degrees, in years-long employment prohibition, removing skilled labor from the workforce.
Claim combines (a) an asserted >90% college-degree rate for H-4 spouses—presumably from ACS/DHS or authors' survey analysis—and (b) immigration policy facts that many H-4 spouses lack work authorization for extended periods; the excerpt does not provide the underlying dataset, sample size, or citations.
Constrained mobility suppresses H-1B wages by 12.2%.
Empirical estimate asserted in the paper (likely from econometric analysis comparing wages under constrained vs. unconstrained mobility); the excerpt does not cite the specific study, dataset, sample size, or methods that produced the 12.2% figure.
Employer-specific sponsorship combined with high switching costs—$5,000+ in fees and multi-year delays—concentrates labor-market power among employers.
Policy/mechanism claim supported by typical filing fee estimates and observed multi-year adjudication/porting constraints; the excerpt does not report a formal empirical test or sample size demonstrating employer market power concentration.
These provisions have generated wait times as extreme as 195 years for Indian nationals in the EB-2 category.
Projection based on visa bulletin/backlog dynamics and issuance rates for EB-2 India; the paper does not show the step-by-step projection or assumptions in the excerpt.
The U.S. employment-based immigration system traps over 1.8 million skilled workers and their families in legal limbo.
Paper's aggregate/backlog calculation presumably using Department of State visa bulletin backlogs, USCIS pending adjustment of status (I-485) inventories, and derivative family counts; the paper does not provide the detailed method or sample breakdown in the excerpt.
Occupational sorting explains a somewhat larger share of the gender gap in Ireland than in other European countries, but a substantial portion remains unexplained, pointing to possible unobserved structural, cultural or organisational factors specific to the Irish labour market.
Decomposition analysis for Ireland using ESJS data showing occupation contributes more to the explained component in Ireland than on average, while the unexplained residual remains large.
Gender gaps are larger and less well explained by observable characteristics among younger cohorts (aged under 35), implying under-representation of women in advanced digital roles is emerging early in careers.
Age-cohort subgroup regressions and decomposition analyses on ESJS data comparing explained/unexplained gaps for workers aged under 35 versus older cohorts.