Evidence (5539 claims)
Adoption
5539 claims
Productivity
4793 claims
Governance
4333 claims
Human-AI Collaboration
3326 claims
Labor Markets
2657 claims
Innovation
2510 claims
Org Design
2469 claims
Skills & Training
2017 claims
Inequality
1378 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 402 | 112 | 67 | 480 | 1076 |
| Governance & Regulation | 402 | 192 | 122 | 62 | 790 |
| Research Productivity | 249 | 98 | 34 | 311 | 697 |
| Organizational Efficiency | 395 | 95 | 70 | 40 | 603 |
| Technology Adoption Rate | 321 | 126 | 73 | 39 | 564 |
| Firm Productivity | 306 | 39 | 70 | 12 | 432 |
| Output Quality | 256 | 66 | 25 | 28 | 375 |
| AI Safety & Ethics | 116 | 177 | 44 | 24 | 363 |
| Market Structure | 107 | 128 | 85 | 14 | 339 |
| Decision Quality | 177 | 76 | 38 | 20 | 315 |
| Fiscal & Macroeconomic | 89 | 58 | 33 | 22 | 209 |
| Employment Level | 77 | 34 | 80 | 9 | 202 |
| Skill Acquisition | 92 | 33 | 40 | 9 | 174 |
| Innovation Output | 120 | 12 | 23 | 12 | 168 |
| Firm Revenue | 98 | 34 | 22 | — | 154 |
| Consumer Welfare | 73 | 31 | 37 | 7 | 148 |
| Task Allocation | 84 | 16 | 33 | 7 | 140 |
| Inequality Measures | 25 | 77 | 32 | 5 | 139 |
| Regulatory Compliance | 54 | 63 | 13 | 3 | 133 |
| Error Rate | 44 | 51 | 6 | — | 101 |
| Task Completion Time | 88 | 5 | 4 | 3 | 100 |
| Training Effectiveness | 58 | 12 | 12 | 16 | 99 |
| Worker Satisfaction | 47 | 32 | 11 | 7 | 97 |
| Wages & Compensation | 53 | 15 | 20 | 5 | 93 |
| Team Performance | 47 | 12 | 15 | 7 | 82 |
| Automation Exposure | 24 | 22 | 9 | 6 | 62 |
| Job Displacement | 6 | 38 | 13 | — | 57 |
| Hiring & Recruitment | 41 | 4 | 6 | 3 | 54 |
| Developer Productivity | 34 | 4 | 3 | 1 | 42 |
| Social Protection | 22 | 10 | 6 | 2 | 40 |
| Creative Output | 16 | 7 | 5 | 1 | 29 |
| Labor Share of Income | 12 | 5 | 9 | — | 26 |
| Skill Obsolescence | 3 | 20 | 2 | — | 25 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
Adoption
Remove filter
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).
AI causes job loss due to the automation of repetitive tasks.
Narrative literature review and synthesis of recent economic studies presented in the paper; no original empirical sample or primary data collection reported.
BT adoption reduces the level of earnings management practice.
Additional empirical tests on the same sample (27,400 firm-years, 2013–2021) comparing firms' earnings management measures before/after or between adopters and non-adopters of BT (earnings management measured by standard accrual-based metrics—details in paper).
The inability of models to reliably self-author useful Skills implies that models typically cannot produce the procedural knowledge they would benefit from consuming.
Interpretation based on the empirical finding that self-generated Skills provided no average benefit; inferred conclusion about model-authored procedural content quality. The paper's claim is supported by the comparative experimental results but the inference about broader capabilities is derived from those results rather than a direct separate measurement.
In some tasks, curated Skills worsened performance: 16 of 84 tasks showed negative deltas.
Per-task delta analysis reported in the paper: authors report 16 tasks with negative deltas where curated Skills reduced pass rate. (Note: the paper elsewhere reports 86 tasks in the benchmark; the negative-task count is reported as 16 of 84 in the paper's per-task summary.)
Developing economies are more vulnerable where employment is concentrated in routine or informal tasks and where reskilling, mobility, and institutional buffers are limited.
Comparative consideration of advanced vs developing economies drawing on macro/sectoral indicators, labor market structure discussions, and existing empirical studies cited conceptually.
Creation of new jobs often lags displacement, producing transitional unemployment and reallocation frictions in the short- to medium-term.
Dynamic/task-based theoretical framing and synthesis of empirical evidence on technology adoption episodes showing delayed job creation relative to displacement.
AI disproportionately automates routine and many middle-skill tasks (both manual and cognitive), displacing corresponding occupations.
Synthesis of occupation- and task-level exposure studies and task-based automation literature referenced in the paper (no new empirical sample provided).
Family- and purpose-driven entrepreneurs (motivated by social stability) experienced larger declines in innovation following income shocks than wealth-driven entrepreneurs.
Subgroup quantitative analysis comparing self-reported post-shock innovation activity across identity-defined groups (family/purpose-driven vs. wealth-driven) within the survey sample; outcome measured conditional on reported income shocks.
Access to digital learning and credential portability could unevenly benefit those with connectivity or prior skills, creating distributional effects and digital divides that should be measured.
Conceptual risk analysis and distributional reasoning based on digital access differentials; no empirical subgroup analysis reported.
Corridor governance is fragmented, with uneven implementation capacity across sending and receiving actors.
Governance gap analysis and desk review of corridor institutional arrangements; qualitative identification of capacity and accountability shortfalls.
Current mandatory pre-departure training is typically delivered late, generically, and with weak assessment, limiting its capacity to change recruitment choices or support migrants after arrival.
Structured desk review of policy and program materials and corridor process mapping identifying timing, actors, and touchpoints; qualitative/administrative evidence rather than quantitative outcome measurement.
Stronger internal corporate governance weakens the AI → executive pay relationship, consistent with governance limiting managerial rent capture during technological change.
Moderation analysis in the paper interacting the firm AI indicator with corporate governance measures; results show a smaller AI effect on pay in firms with stronger governance (same sample and regression framework).
Traditional extrapolation-based employment forecasting (as used in current BLS/standard practice) is inadequate for capturing AI-driven labor market change.
Conceptual argument in the paper highlighting limitations of extrapolation methods (failure to distinguish automation vs augmentation, inability to capture rapid nonlinear adoption dynamics and demographic heterogeneity). No empirical test or sample is reported; critique is supported by theoretical considerations and examples rather than an applied dataset.
Inflation and geopolitical fragmentation can raise the cost of AI deployment (hardware shortages, supply constraints) and complicate cross-border data flows, slowing diffusion or creating regionalized AI ecosystems.
Conceptual argument linking macroeconomic and geopolitical constraints to AI deployment costs; no empirical cost-accounting or cross-country diffusion analysis provided in the paper.
Mandel's account—that capitalist production relations, class struggle, and global imbalances shape the course and consequences of waves—implies that crises expose and amplify supply-chain fragilities and bargaining conflicts that affect profitability.
Theoretical interpretation of Mandel's political-economy literature and historical examples (qualitative).
Platforms optimized for engagement can produce externalities that distort lived temporality (loss of presence and meaning) beyond standard attention‑capture harms.
Argument synthesizing platform literature and phenomenological concerns; no new empirical analysis of platform effects provided.
Contemporary transhumanist and neurotechnology developments (BCIs, neural digital twins, human–AI collaboration) have advanced technologically but lack a robust conceptual core focused on lived experience and temporality.
Survey and synthesis of existing literatures reported in the paper (conceptual review); no systematic empirical content analysis or coded sample size provided.
High PIGRS scores associate with genomic instability (higher tumor mutational burden and MATH heterogeneity scores) and immune‑escape signatures.
Association analyses within the PIGRS study linking high risk scores to higher TMB, elevated MATH scores, and immune evasion markers (multi‑omics and immune gene set analyses reported).
The price-of-transparency quantifies how increased observability (e.g., from disclosure or regulation) can reduce the effectiveness of deception-based defenses, informing policy tradeoffs.
Formal definition of price of transparency and analytical results showing its effect; policy implication drawn in discussion (theoretical analysis, no empirical policy case studies).
High upfront and maintenance costs create scale advantages for larger institutions or centralized providers, potentially concentrating market power among well-resourced curriculum developers.
Economic inference from cost structure described in paper; no market concentration empirical data provided.
Disadvantages and risks include significant resource investment, complexity aligning multiple standards, and a high demand for continuous updates and audits.
Paper's risks section (author assertion); no quantified cost or burden data.
Implementing this program requires substantial resources and ongoing governance.
Author assertions in disadvantages/risks section; no cost accounting or empirical costing data provided.
One-size-fits-all AI competency approaches fail to account for local labor markets, pedagogical traditions, and resource realities; respondents favor context-aware frameworks allowing discipline-specific adaptation.
Thematic analysis of open-ended responses expressing preferences for context-aware, flexible frameworks; survey items mapped to UNESCO competency frameworks asking about adaptability and local relevance.
Infrastructural limitations (bandwidth, computing resources, licensing costs) disproportionately affect respondents in the Global South and smaller institutions.
Comparative descriptive analysis by region (Global South vs Global North) and institution size/type within the >600 respondent sample; survey items on infrastructure and costs; thematic coding supporting differential impact.
Practical barriers—software access, available datasets, and lab time—limit experiential learning that builds AI competency.
Survey items listing barriers to AI learning and training; thematic coding of open responses highlighting software, dataset, and scheduling constraints.
Respondents cite limited opportunities for applied, project-based learning with AI tools; where AI appears in curricula, coverage is more theory-oriented than hands-on.
Quantitative items and open-ended responses about types of training and curricular integration; thematic analysis of qualitative data indicating prevalence of theory-focused instruction versus applied opportunities.
Many institutions lack clear, consistent, or context-sensitive policies for AI use in learning, assessment, and academic integrity.
Survey questions about the presence and clarity of institutional AI policies and thematic coding of open-ended responses reporting policy gaps; descriptive summaries across respondents.
Educators frequently report lower confidence in teaching AI-relevant skills than students report in using AI tools, reducing instructional capacity.
Survey items measuring self-reported competency/confidence for educators (teaching) and students (using); comparative descriptive analysis across roles within the >600 participant sample.
Proprietary models trained on large clinical datasets can create high entry barriers, concentrating market power among a few platform firms and increasing prices for hospitals.
Market-structure and platform economics analysis in the paper; empirical evidence of concentration in GenAI healthcare is limited and no firm-level market-share data are provided.
Liability and accountability gaps exist for AI-suggested errors: it is unclear whether vendors, hospitals, or clinicians are responsible for harms resulting from GenAI CDS recommendations.
Policy and legal analysis discussed in the paper; this is a structural/legal observation rather than an empirical finding and no case-law sample size is provided.
Current simulation practice is insufficiently integrated with enabling technologies (digital twins, data analytics, AI/ML) and with relevant government policy constraints.
Synthesis of literature and gap analysis in the paper; assertions are conceptual and not empirically tested within the paper.
Current simulation practice has limited strategic orientation, often focusing more on tactical and operational questions than on firm strategy.
Literature review and analysis in the paper highlighting the emphasis in existing studies on tactical/operational problems.
Current simulation practice lacks contextualization to firm‑ and industry‑specific realities.
Findings from the paper's literature review and critique sections; no new empirical measurement provided.
Current manufacturing and supply‑chain simulation practices are insufficiently contextualized, strategically focused, or integrated with modern technologies and policy considerations.
Literature review and critique of existing simulation practice presented in the paper; no original empirical data or case studies.
AI and platform integration can increase systemic interconnectedness and winner-take-all dynamics, raising systemic-risk concerns.
Theoretical discussion and policy-oriented literature review recommending macroprudential incorporation of algorithmic concentration and network effects; no quantitative systemic-risk model results provided in the abstract.
Regulatory gaps, fragmentation across providers, and weak governance of data/AI pose risks to financial stability, consumer protection, and trust.
Policy and literature review identifying documented regulatory lacunae and governance risks; supported by qualitative case examples rather than quantified systemic risk metrics in the paper summary.
ML-based IDS models are vulnerable to adversarial examples, poisoning attacks, and evasion techniques, raising security and robustness concerns.
Survey references and synthesis of works discussing/adapting adversarial attacks and poisoning against ML models in network/IoT contexts.
Heterogeneity of devices, protocols, and feature sets complicates generalization of IDS models across different IoT environments.
Literature reports limited cross-device generalization and difficulties transferring models between device types; survey highlights heterogeneity as a major barrier.
Practical constraints — device heterogeneity, resource limits, dataset shortcomings, and ML pipeline pitfalls — prevent many research models from reaching operational use.
Thematic analysis across surveyed studies highlighting recurring barriers: heterogeneous device/protocol stacks, limited compute/memory on edge devices, dataset limitations, and methodological pitfalls.
Personalization raises distributional concerns and risks of manipulation or biased treatment; regulators may need to set transparency, fairness, and data-use standards.
Policy analysis and normative recommendation based on known risks of personalization systems; not empirically demonstrated in robotic deployments here.
LLM-based personalization generates context-aware responses but often fails to model long-term preferences and fine-grained user/item relations needed for consistent, proactive personalization.
Conceptual critique based on surveyed limitations of LLM-based approaches; no new experimental data reported.
Value-based pricing remains underdeveloped in practice because theory and empirical evidence are fragmented and sparse.
Synthesis from the SLR showing fragmented theoretical approaches and empirical gaps across the 30 included studies; authors' interpretation in discussion.
Rural digital divides mean AI benefits will be unevenly distributed; models trained on digitally-rich urban records could bias resource allocation away from rural trainees.
Analytical/risk assessment in the paper noting distributional risks; no empirical bias measurement presented.