Evidence (13870 claims)
Adoption
8467 claims
Productivity
7558 claims
Governance
6805 claims
Human-AI Collaboration
6363 claims
Org Design
4132 claims
Innovation
4065 claims
Labor Markets
3526 claims
Skills & Training
2945 claims
Inequality
2066 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 749 | 196 | 98 | 892 | 1984 |
| Governance & Regulation | 817 | 394 | 188 | 121 | 1544 |
| Organizational Efficiency | 771 | 189 | 124 | 83 | 1177 |
| Technology Adoption Rate | 627 | 233 | 123 | 96 | 1088 |
| Research Productivity | 411 | 123 | 56 | 332 | 933 |
| Output Quality | 467 | 178 | 59 | 47 | 751 |
| Decision Quality | 320 | 174 | 75 | 42 | 618 |
| Firm Productivity | 435 | 55 | 88 | 20 | 604 |
| AI Safety & Ethics | 214 | 276 | 65 | 33 | 593 |
| Market Structure | 178 | 167 | 122 | 24 | 496 |
| Task Allocation | 207 | 64 | 71 | 32 | 379 |
| Skill Acquisition | 165 | 59 | 60 | 17 | 301 |
| Innovation Output | 203 | 27 | 43 | 18 | 292 |
| Employment Level | 105 | 52 | 107 | 13 | 279 |
| Fiscal & Macroeconomic | 131 | 69 | 43 | 26 | 276 |
| Consumer Welfare | 116 | 63 | 42 | 11 | 232 |
| Firm Revenue | 150 | 48 | 26 | 3 | 227 |
| Inequality Measures | 44 | 122 | 49 | 6 | 221 |
| Task Completion Time | 169 | 29 | 8 | 12 | 219 |
| Worker Satisfaction | 89 | 63 | 20 | 12 | 184 |
| Error Rate | 69 | 92 | 10 | 2 | 173 |
| Regulatory Compliance | 76 | 68 | 14 | 5 | 163 |
| Training Effectiveness | 93 | 21 | 13 | 19 | 148 |
| Wages & Compensation | 77 | 36 | 25 | 6 | 144 |
| Automation Exposure | 51 | 54 | 22 | 12 | 142 |
| Team Performance | 86 | 17 | 27 | 9 | 140 |
| Developer Productivity | 94 | 17 | 14 | 6 | 132 |
| Job Displacement | 12 | 80 | 20 | 1 | 113 |
| Hiring & Recruitment | 51 | 7 | 8 | 3 | 69 |
| Creative Output | 31 | 17 | 7 | 3 | 59 |
| Skill Obsolescence | 5 | 46 | 6 | 1 | 58 |
| Social Protection | 27 | 16 | 8 | 2 | 53 |
| Labor Share of Income | 17 | 17 | 17 | — | 51 |
| Worker Turnover | 11 | 12 | — | 3 | 26 |
| Industry | — | — | — | 1 | 1 |
The paper discusses a regulatory framework for token futures markets, providing a theoretical foundation and practical roadmap for the financialization of compute resources.
Policy/regulatory discussion and recommendations included in the paper; draws on comparisons to existing commodity regulation and futures markets.
The paper explores the feasibility of GPU compute futures as an alternative or complement to token futures.
Discussion/feasibility analysis in the paper (conceptual and comparative discussion; not presented as empirical field evidence).
Simulation results show that, under an application-layer demand explosion scenario, token futures can reduce enterprise compute cost volatility by 62%–78%.
Monte Carlo simulation results based on the constructed mean-reverting jump-diffusion stochastic process model; scenario described as 'application-layer demand explosion'. (No numerical sample size reported in the abstract.)
The authors propose a complete design for standardized token futures contracts, including the definition of a Standard Inference Token (SIT), contract specifications, settlement mechanisms, margin systems, and market-maker regimes.
Normative/proposal section of the paper specifying contract design components and market microstructure recommendations.
Tokens consumed by AI inference are evolving into a new type of commodity.
Conceptual/systematic analysis and argumentation presented in the paper (comparisons to established commodities and discussion of commodity attributes).
By enabling developers without initial capital to participate in the digital economy, RSI could unlock the 'latent jobs dividend' in low-income countries and help address local challenges in health, agriculture, and services.
Societal-impact argument in the paper linking the RSI model to potential employment gains and localized solutions; speculative extrapolation, no empirical employment estimates or pilot studies reported.
The RSI model could stimulate innovation in the ecosystem.
Argument based on lowered financial barriers and incentive structures from the paper's theoretical comparative analysis; no empirical measures of innovation provided.
The RSI model aligns stakeholder interests (platforms and developers).
Theoretical argument and incentive-alignment reasoning in the paper's comparative framework; no empirical validation presented.
A comparative analysis in the paper shows that the RSI model lowers entry barriers for developers.
Detailed comparative (theoretical) analysis within the paper contrasting existing models and RSI; no empirical trial, sample, or randomized test reported.
Generative AI platforms (Google AI Studio, OpenAI, Anthropic) provide infrastructures (APIs, models) that are transforming the application development ecosystem.
Statement in paper based on literature review and descriptive framing of current platforms; no empirical sample or quantitative test reported.
Policy should address not only the aftermath of AI labor displacement but also the competitive incentives that drive it.
Normative implication drawn from the model's findings; recommendation in the paper's conclusion based on theoretical results.
Only a Pigouvian automation tax can eliminate the excess automation in the model.
Theoretical welfare analysis demonstrating that a properly set Pigouvian tax that internalizes the demand externality restores the socially optimal level of automation in the model; analytical result, no empirical sample.
The paper proposes a dual-nudge governance architecture leveraging the DHDE to redistribute cross-prefectural flows and reduce economic leakage.
Policy/design proposal presented by the authors as an outcome of the DHDE adaptation and analysis (conceptual/proposed intervention).
The AI-driven decision support system achieves out-of-sample predictive performance of 68% (R^2 = 0.683).
Model performance metric reported in the paper (out-of-sample R^2 value); presumably from held-out validation or cross-validation on the datasets.
The AI-driven decision support system achieves in-sample explanatory power of 81% (R^2 = 0.810).
Model performance metric reported in the paper (in-sample R^2 value); derived from applying the DSS to the supplied datasets.
Financial digital intelligence enhances innovation by strengthening regional industry–university–research collaboration.
Authors report this channel from mechanism/mediation tests using the same empirical sample (5,731 observations, 2015–2022); specific measures of collaboration or identification strategy not provided in excerpt.
Financial digital intelligence enhances innovation by reducing transaction costs.
Mechanism analysis reported by authors on the same panel dataset (5,731 observations, 2015–2022); reduction in transaction costs is presented as a mediating channel (details of measurement/identification not included in excerpt).
Financial digital intelligence enhances innovation by improving corporate information disclosure.
Mechanism analysis reported in paper using same empirical sample (5,731 observations, 2015–2022); authors identify corporate information disclosure as a mediating channel (specific identification strategy not provided in excerpt).
Financial digital intelligence remarkably boosts the innovative development of strategic emerging industries.
Empirical analysis using panel data from 2015–2022 comprising 5,731 observations covering 789 listed companies and 114 prefecture-level cities in China (methods not specified in excerpt; presumably regression analysis on firm/city-level panel).
In production, the system received high satisfaction from both domain experts and developers, with all participants reporting full satisfaction with communication efficiency.
Post-deployment user feedback / satisfaction reports mentioned in paper (no numeric participant count provided).
The automated workflow saved an estimated 979 engineering hours.
Aggregate time-savings estimate reported in paper (derived from per-API time reduction × number of APIs).
The automated workflow reduces per-API development time from approximately 5 hours to under 7 minutes.
Time-per-API comparison reported in paper based on evaluation on spapi (comparison of manual vs automated per-API time).
The automated workflow achieves 93.7% F1 score.
Empirical evaluation on spapi (F1 reported); presumably computed over the evaluated API items/endpoints.
We address this gap through a graph-based workflow optimization approach that progressively replaces manual coordination with LLM-powered services, enabling incremental adoption without disrupting established practices.
Description of proposed method (graph-based workflow + LLM-powered services) and claim of design enabling incremental adoption; supported by subsequent case evaluation.
The work underscores the urgency of tangible actions aimed at closing the AI divide and allowing Africa to actively shape its AI future.
Concluding normative claim in the paper, supported by the paper's synthesis of identified infrastructural and policy barriers and the illustrative ACT tool.
We introduce the Africa AI Compute Tracker (ACT), an interactive map to monitor the availability of AI-ready HPC systems throughout the continent.
Paper reports development and introduction of the ACT tool; the claim is about the authors' own deliverable (an interactive map consolidating HPC availability data).
Sustainable AI adoption requires robust digital foundations through balanced access to compute, data, and the energy that makes it possible (the 'right enablers').
Normative claim grounded in the paper's stated quantitative and qualitative analysis and synthesis of official declarations; presented as a central conceptual conclusion.
Organizational size moderates the adoption–efficiency relationship such that larger firms realize proportionally greater efficiency gains from AI adoption.
Reported moderation effect in the PLS-PM analysis testing organizational size as a moderator of the relationship between AI adoption and recruitment efficiency metrics across sampled organizations.
Procedural fairness perceptions positively predict employee experience outcomes, including organizational commitment, job satisfaction, and employer trust.
PLS-PM paths from procedural fairness perceptions to employee experience measures (organizational commitment, job satisfaction, employer trust) using survey data from HR professionals' reports.
Algorithmic transparency is a strong predictor of procedural fairness perceptions.
PLS-PM results linking measured algorithmic transparency to procedural fairness perceptions in the survey data (n=523 respondents).
AI adoption is positively associated with improvements in quality-of-hire.
PLS-PM association between AI adoption and reported quality-of-hire improvement from HR respondents across sampled organizations.
AI adoption is positively associated with reductions in cost-per-hire.
PLS-PM association between AI adoption and cost-per-hire reduction reported in the survey (firm-level outcomes across sampled organizations).
AI adoption is positively associated with reductions in time-to-hire (recruitment time).
PLS-PM association between AI adoption and recruitment efficiency metrics reported in the survey (firm-level outcomes across sampled organizations).
Top management support and HR digital readiness are both positively associated with organizational AI adoption, with top management support demonstrating greater explanatory power.
PLS-PM tests of organizational antecedents predicting organizational AI adoption using survey responses aggregated to organization level (184 organizations referenced).
Perceived usefulness and perceived ease of use significantly predict AI adoption intention, with perceived usefulness exhibiting a stronger effect.
PLS-PM results on relationships between TAM constructs (perceived usefulness, perceived ease of use) and adoption intention using survey data (n=523).
A large portion of the interactive activities' AI market value (26%) involves transferring information.
Descriptive subcategory statistic: within interactive activities, authors report 26% of market value pertains to information transfer tasks.
Interactive activities (which include both information-based and physical activities) account for 48% of AI market value.
Descriptive aggregate: authors define an 'interactive' category spanning info and physical activities and report it holds 48% of AI market value.
A substantial portion of AI market value (36%) is used in activities that involve creating information.
Descriptive aggregate: subcategory within information-based activities—authors report 36% of market value allocated to 'creating information'.
Most of the AI market value is used in information-based activities (72%).
Descriptive aggregate: authors categorize activities into information-based vs physical and report that 72% of estimated AI market value maps to information-based activities.
There is a highly uneven distribution of AI market value across activities: the top 1.6% of activities account for over 60% of AI market value.
Descriptive statistical result from mapping estimated AI market values to the ~20K activities; authors report concentration metrics (top 1.6% share >60%).
We use the data about AI software and robotic systems to generate graphical displays of how the estimated units and market values of all worldwide AI systems used today are distributed across the work activities that these systems help perform.
Analytic/mapping procedure: authors combine classifications of software (13,275) and robots (20.8M) with market-value estimates to create visual distributions across activities.
We classify a worldwide tally of 20.8 million robotic systems using the developed work-activity ontology.
Empirical classification/counting: authors report mapping 20.8 million robotic systems worldwide to the activity ontology.
We classify descriptions of 13,275 AI software applications using the developed work-activity ontology.
Empirical classification: authors state they mapped 13,275 AI software application descriptions to the ontology.
We disaggregate and then substantially reorganize the approximately 20K activities in the US Department of Labor's O*NET occupational database to produce a comprehensive ontology of work activities.
Methodological: authors report transforming the O*NET activity taxonomy (~20,000 activity-level records) by disaggregation and reorganization into a new ontology.
Models trained in EnterpriseLab remain robust across diverse enterprise benchmarks, including EnterpriseBench (+10%) and CRMArena (+10%).
Benchmark evaluations reported in the paper showing reported +10% improvements on EnterpriseBench and CRMArena relative to baseline; exact baselines, statistical tests, and sample sizes are not specified in the abstract.
8B-parameter models trained in EnterpriseLab reduce inference costs by 8-10x compared to frontier models (implied GPT-4o).
Empirical cost comparison reported in the paper; the abstract states an 8-10x reduction in inference costs for the 8B models trained in EnterpriseLab versus the referenced frontier model(s). Detailed cost accounting and sample sizes not provided in the abstract.
8B-parameter models trained within EnterpriseLab match GPT-4o's performance on complex enterprise workflows.
Empirical evaluation reported in the paper comparing 8B-parameter models trained in EnterpriseLab to GPT-4o on complex enterprise workflows; specific benchmark tests and metrics are referenced but details (sample sizes, exact metrics) are not provided in the abstract.
We validate the platform through EnterpriseArena, an instantiation with 15 applications and 140+ tools across IT, HR, sales, and engineering domains.
Reported instantiation/experimental setup in the paper: EnterpriseArena contains 15 applications and 140+ tools spanning specified domains.
EnterpriseLab provides integrated training pipelines with continuous evaluation.
System/design claim in paper describing integrated training and evaluation tooling as part of the platform.
EnterpriseLab includes automated trajectory synthesis that programmatically generates training data from environment schemas.
System/design claim described in paper; supported by the authors' description of an automated data-generation component.