Evidence (5267 claims)
Adoption
5267 claims
Productivity
4560 claims
Governance
4137 claims
Human-AI Collaboration
3103 claims
Labor Markets
2506 claims
Innovation
2354 claims
Org Design
2340 claims
Skills & Training
1945 claims
Inequality
1322 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 378 | 106 | 59 | 455 | 1007 |
| Governance & Regulation | 379 | 176 | 116 | 58 | 739 |
| Research Productivity | 240 | 96 | 34 | 294 | 668 |
| Organizational Efficiency | 370 | 82 | 63 | 35 | 553 |
| Technology Adoption Rate | 296 | 118 | 66 | 29 | 513 |
| Firm Productivity | 277 | 34 | 68 | 10 | 394 |
| AI Safety & Ethics | 117 | 177 | 44 | 24 | 364 |
| Output Quality | 244 | 61 | 23 | 26 | 354 |
| Market Structure | 107 | 123 | 85 | 14 | 334 |
| Decision Quality | 168 | 74 | 37 | 19 | 301 |
| Fiscal & Macroeconomic | 75 | 52 | 32 | 21 | 187 |
| Employment Level | 70 | 32 | 74 | 8 | 186 |
| Skill Acquisition | 89 | 32 | 39 | 9 | 169 |
| Firm Revenue | 96 | 34 | 22 | — | 152 |
| Innovation Output | 106 | 12 | 21 | 11 | 151 |
| Consumer Welfare | 70 | 30 | 37 | 7 | 144 |
| Regulatory Compliance | 52 | 61 | 13 | 3 | 129 |
| Inequality Measures | 24 | 68 | 31 | 4 | 127 |
| Task Allocation | 75 | 11 | 29 | 6 | 121 |
| Training Effectiveness | 55 | 12 | 12 | 16 | 96 |
| Error Rate | 42 | 48 | 6 | — | 96 |
| Worker Satisfaction | 45 | 32 | 11 | 6 | 94 |
| Task Completion Time | 78 | 5 | 4 | 2 | 89 |
| Wages & Compensation | 46 | 13 | 19 | 5 | 83 |
| Team Performance | 44 | 9 | 15 | 7 | 76 |
| Hiring & Recruitment | 39 | 4 | 6 | 3 | 52 |
| Automation Exposure | 18 | 17 | 9 | 5 | 50 |
| Job Displacement | 5 | 31 | 12 | — | 48 |
| Social Protection | 21 | 10 | 6 | 2 | 39 |
| Developer Productivity | 29 | 3 | 3 | 1 | 36 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
| Skill Obsolescence | 3 | 19 | 2 | — | 24 |
| Creative Output | 15 | 5 | 3 | 1 | 24 |
| Labor Share of Income | 10 | 4 | 9 | — | 23 |
Adoption
Remove filter
This paper investigates societal applications of AI across domains such as healthcare, education, accessibility, environmental management, emergency response, and civic administration.
Descriptive statement of the paper's scope and methods (literature review / cross-domain analysis implied); the abstract lists the domains but does not specify empirical procedures or sample sizes.
Chatbot suggestions were artificially varied in aggregate accuracy across treatment conditions from low (53%) to high (100%).
Paper describes experimental manipulation of chatbot suggestion accuracy with aggregate accuracies ranging from 53% to 100%; manipulation method (how suggestions were generated or sampled) described in methods (not fully detailed in excerpt).
Caseworkers in the control condition (no chatbot suggestions) had a mean accuracy of 49%.
Reported experimental outcome: mean accuracy for control group = 49%; based on the randomized experiment using the 770-question benchmark.
We conducted a randomized experiment with caseworkers recruited from nonprofit outreach organizations in Los Angeles.
Paper describes a randomized experiment recruiting caseworkers from nonprofit outreach organizations in Los Angeles; sample size and recruitment details not given in the excerpt.
The benchmark questions have corresponding expert-verified answers.
Paper states benchmark questions have expert-verified answers; verification method and number/credentials of experts not specified in the excerpt.
We created a 770-question multiple-choice benchmark dataset of difficult, but realistic questions that a caseworker might receive.
Paper reports creation of a benchmark dataset containing 770 multiple-choice questions described as difficult and realistic; questions and dataset construction described in methods (no sample-of-questions or external validation details provided in the excerpt).
The study's conclusions draw on three complementary evidence bases: (a) task-level evidence on what generative AI can already do in practice; (b) occupational exposure and complementarity analysis using Philippine labor force data; and (c) firm- and worker-level evidence on AI adoption.
Description of methods and data sources in the paper: task-level capability testing/assessment, analysis of national labor force/occupation data for exposure/complementarity, and firm/worker surveys or qualitative adoption evidence.
There is a need for more longitudinal and cross-country studies to better understand the long-term value creation of ERM in MSMEs.
Authors' conclusion and identified research gaps based on the scope and limitations of the existing literature reviewed (i.e., predominance of cross-sectional or single-country studies).
Extensive experiments were conducted using both synthetic and real hospital datasets to evaluate the framework.
Statement in the paper indicating experiments on synthetic and real datasets; exact sizes, sources, and composition of these datasets are not provided in the excerpt.
The paper explains the main legal frameworks that currently regulate AI in India, as well as proposals for future legislation.
Author's legal and policy analysis / document review of existing statutes and proposed laws (qualitative review). No quantitative sample size; based on review of legal texts and policy proposals cited in the article.
DDDM was quantified using AI language models, specifically BERT and ChatGLM2-6B.
Methodological description in the paper stating that BERT and ChatGLM2-6B were leveraged to quantify the extent of DDDM (implementation details, training/data specifics, and sample not provided in the excerpt).
A “macro approach” that (1) directly models equilibrium behavior of large employers, (2) combines macro data with empirical estimates of employers’ responses (from the micro approach) to estimate the model, and (3) uses the model to compute aggregate costs of monopsony and optimal policies, is the appropriate methodological response.
Methodological proposal set out by the paper; this is a description of the authors' recommended empirical/theoretical strategy rather than an empirical finding. The excerpt contains no implementation details, datasets, or estimation results.
The traditional theoretical and empirical “micro approach” to studying labor market power requires that firms are small and atomistic.
Conceptual/theoretical characterization of the micro approach stated by the paper; no empirical sample, dataset, or formal model provided in the excerpt.
Most evidence came from retrospective studies or meta-analyses, with limited prospective or randomized controlled trials.
Summary of study designs across the 40 included studies as reported in the review.
The impact of AI on patient outcomes (e.g., mortality, rebleeding) was rarely addressed.
Statement in results indicating few included studies reported patient-centered outcomes such as mortality or rebleeding.
This systematic review adhered to PRISMA 2020 guidelines.
Methods statement in the paper specifying adherence to PRISMA 2020; the review included 40 studies.
The review focuses on AI applications within small‑scale business environments, with a special focus on women‑owned micro firms in Jaipur, India.
Scope and aim articulated in the paper; geographic and demographic focus explicitly stated by the authors.
The systematic review follows PRISMA 2020 guidelines.
Methodological statement in the paper indicating adherence to PRISMA 2020 for the review process.
After screening and eligibility filtering, 55 open‑access journal articles were included for in‑depth analysis.
PRISMA‑guided screening and eligibility process reported in the review; final included sample explicitly stated as 55 open‑access journal articles.
A Scopus search identified 265 records using keywords related to women’s entrepreneurship and AI.
Systematic literature search reported in the paper following PRISMA 2020; search executed in Scopus with specified keywords; initial yield stated as 265 records.
This research examined three countries (China, the United States, and Germany) using panel vector autoregressive (panel VAR) and difference-in-differences (DID) methods to assess how technology and public policy interventions affect emissions reductions.
Study design reported in the paper: sample of three countries (China, US, Germany) and application of panel VAR and DID methods; specific time period and sample size not provided in the summary.
Social assistance (SA) is defined here as noncontributory social transfers (including cash, vouchers, or in-kind transfers to families or individuals, including the elderly), public works programs, fee waivers, and subsidies.
Explicit definitional statement in the introduction (authors' operational definition for the chapter).
This chapter focuses on low- and middle-income countries (LMICs) and uses a 'review of reviews' approach to summarize the policy discourse and evidence on social protection and gender in adulthood, concentrating on social assistance, social care, and social insurance.
Methodological and scope statement explicitly given in the introduction (author-declared approach and focus).
Viable transition pathways are operationally defined in this study as sharing at least 3 skills and achieving at least 50% skill transfer.
Methodological definition stated in the paper used to determine whether a job-to-job transition is considered viable.
We identified 4,534 feasible transitions between jobs in the dataset.
Count of feasible job-to-job transition pairs found in the knowledge graph analysis (4,534 transitions reported).
We constructed and validated a knowledge graph of 9,978 Egyptian job postings, 19,766 skill activities, and 84,346 job-skill relationships with a 0.74% error rate.
Empirical construction and validation of a knowledge graph using a dataset of 9,978 job postings, 19,766 distinct skill/activity nodes, and 84,346 job–skill edges; reported overall error rate 0.74% (validation method not detailed in the excerpt).
The framework was evaluated on 2,847 queries across 15 task categories.
Paper reports an evaluation dataset consisting of 2,847 queries spanning 15 task categories; used as the sample for reported empirical results.
Non-text processing paths use SLM-assisted modality decomposition.
Paper reports that non-text queries are decomposed using SLM-assisted modality decomposition; described as the non-text routing approach in the framework.
For text-only queries, the framework uses learned routing via RouteLLM.
Paper states text-only routing is handled by a learned model named RouteLLM; presented as part of the system architecture.
A central Supervisor dynamically decomposes user queries, delegates subtasks to modality-appropriate tools (e.g., object detection, OCR, speech transcription), and synthesizes results through adaptive routing strategies rather than predetermined decision trees.
Methodological description in the paper of a Supervisor component that performs dynamic decomposition, delegation to modality-appropriate tools (examples given), and adaptive routing; supported by the framework's implementation details.
We present an agentic AI framework for autonomous multimodal query processing that coordinates specialized tools across text, image, audio, video, and document modalities.
Paper describes the framework design and components (Supervisor, modality-specific tools) and states support for text, image, audio, video, and document modalities; no external benchmark cited for this capability beyond the paper's own implementation.
The study employs an input–output (I–O) modeling framework using IMPLAN 2022 data to estimate direct, indirect, and induced impacts of investments in greenhouse and robotics sectors for Northwest Indiana as part of Project TRAVERSE.
Explicit methodological statement in the paper: use of IMPLAN 2022 I–O model; geographic scope NWI; linkage to EDA Project TRAVERSE.
We extract the Big 5 personality traits from facial images of 96,000 MBA graduates using advances in AI and LinkedIn microdata.
Methodological claim reported in the paper: AI-based model applied to facial images linked to LinkedIn microdata for a sample of 96,000 MBA graduates; extraction yields 'Photo Big 5' trait scores.
The essay reviews seven books from the past dozen years by social scientists examining the economic impact of artificial intelligence (AI).
Qualitative book-review performed by the author; sample size explicitly stated as seven books published within the last ~12 years; method = synthesis/assessment of those seven books.
This systematic review follows PRISMA guidelines to examine the evolution, advancements, and state-of-the-art AI applications for GS-BESS optimization.
Methodological statement in the paper indicating the use of PRISMA guidelines for the review process. The excerpt does not include the PRISMA flow diagram or the exact article selection numbers.
The study is limited by the scope of available industry data and the generalisability of case study findings.
Explicit limitation reported in the paper summary stating constraints related to industry data availability and generalisability of case studies.
The research adopts a mixed-method approach, combining theoretical analysis with empirical insights, and uses data gathered from the 'AI-driven transformation' Scopus database.
Explicit methodological statement in the paper summary: mixed-method design and Scopus database as the data source. (No further methodological details or sample counts provided in the summary.)
The conceptual model for the study is grounded in the Resource-Based View (RBV) and the Technology-Organization-Environment (TOE) framework.
Theory section of the paper: model development explicitly references RBV and TOE as theoretical foundations for selecting determinants and mediators.
The data were analysed using partial least squares structural equation modeling (PLS-SEM).
Methods section: PLS-SEM specified as the primary analytical technique for hypothesis testing and mediation analysis.
Data were collected via a cross-sectional survey of 312 senior managers across diverse UK industries.
Study methods: described sample = 312 senior managers from multiple UK industries; cross-sectional survey instrument and sampling reported in methods section.
The experimental sample underlying the statistical tests comprised 20 observations (implied by ANOVA degrees of freedom: df between = 1, df within = 18).
Interpretation of the reported one-way ANOVA degrees of freedom (F(1,18) for multiple outcomes) indicating total N = 20 observations.
Field experiments at the Al‐Ra'id Research Station in Baghdad during the 2025 season compared conventional diesel‐based irrigation with AI‐assisted irrigation using soil moisture sensors, IoT controllers, and predictive weather algorithms.
Reported field experiment design in the paper (Al‐Ra'id Research Station, Baghdad, 2025 season) specifying two treatments: conventional diesel irrigation vs AI-assisted irrigation using soil moisture sensors, IoT controllers, and predictive weather algorithms.
Definitions and scopes of Material Passports vary among authors.
Content analysis of the 46 included studies showing differing definitions and scope treatments for MPs reported by the authors.
Among the included studies, 65% focused primarily on Material Passports (MPs), while 35% addressed MPs within the broader context of a circular economy (CE).
Quantitative categorization of the 46 included studies reported in the paper (percentages attributed to focus areas).
A total of 54 peer-reviewed articles and book chapters were screened from the Scopus database, of which 46 were included for in-depth analysis in April 2025.
Reported screening and inclusion counts from the Scopus search (54 screened, 46 included); date of in-depth analysis given as April 2025.
This article presents a Systematic Literature Review (SLR) following the PRISMA methodology.
Stated methodology in the paper: SLR using PRISMA; literature search performed in Scopus; review process and inclusion/exclusion described (screening and inclusion counts reported).
Future research could strengthen causal identification by exploiting exogenous policy shocks rather than relying solely on matching methods like PSM.
Authors' methodological suggestion for future work, based on limitations of current causal inference strategy (PSM and observational panel regression).
Propensity Score Matching (PSM) and other robustness checks were used to mitigate selection bias and support the causal interpretation of AI's effects.
Paper reports use of Propensity Score Matching in robustness analyses on the panel of A-share-listed design firms (2014–2023).
The paper operationalizes firm-level AI exposure by constructing an AI lexicon via natural language processing and applying text analysis to annual reports and patents to generate enterprise-level AI indicators.
Described methodology: NLP to generate an AI lexicon and text-analysis of annual reports and patents to build AI measures for each listed design enterprise in the 2014–2023 panel.
By integrating dynamic capabilities theory with a micro foundations perspective, the study proposes a conditional model that reframes the essential challenge from technology adoption to organizational adaptation.
Model/theory construction presented in the paper (conceptual integration). This is a methodological/theoretical claim about the paper's contribution; no empirical validation provided.