Evidence (16496 claims)
Search and filter individual claims pulled from the papers. Looking for a specific finding ("what's the effect on wages?"), you're in the right place. Want to compare whole outcome categories against each other instead? Use the Evidence Explorer.
The board below groups claims two ways: by broad theme (nine paper-level topics) and by outcome category (the 34 claim-level outcomes that the Explorer and Syntheses also use).
Browse by theme
Nine broad, paper-level topics. Click one to filter the claims below.
Adoption
9875 claims
Filter claims →
Productivity
8807 claims
Filter claims →
Governance
7870 claims
Filter claims →
Human-AI Collaboration
7560 claims
Filter claims →
Org Design
4892 claims
Filter claims →
Innovation
4781 claims
Filter claims →
Labor Markets
4004 claims
Filter claims →
Skills & Training
3308 claims
Filter claims →
Inequality
2332 claims
Filter claims →
Claims by outcome category
Counts by direction of finding. These are the same 34 outcome categories the Explorer compares and the Syntheses are written for. A linked row has a published synthesis.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 870 | 233 | 116 | 1066 | 2363 |
| Governance & Regulation | 976 | 451 | 218 | 133 | 1809 |
| Organizational Efficiency | 949 | 224 | 144 | 88 | 1416 |
| Technology Adoption Rate | 764 | 287 | 141 | 122 | 1325 |
| Research Productivity | 501 | 152 | 74 | 362 | 1101 |
| Output Quality | 542 | 216 | 69 | 69 | 896 |
| Decision Quality | 387 | 198 | 94 | 54 | 740 |
| Firm Productivity | 513 | 67 | 101 | 27 | 714 |
| AI Safety & Ethics | 249 | 303 | 73 | 36 | 667 |
| Market Structure | 190 | 192 | 134 | 27 | 548 |
| Task Allocation | 243 | 77 | 91 | 36 | 452 |
| Innovation Output | 291 | 33 | 55 | 20 | 401 |
| Skill Acquisition | 206 | 72 | 65 | 21 | 364 |
| Employment Level | 133 | 63 | 115 | 22 | 335 |
| Fiscal & Macroeconomic | 153 | 79 | 52 | 32 | 323 |
| Task Completion Time | 206 | 37 | 12 | 15 | 272 |
| Firm Revenue | 179 | 52 | 29 | 5 | 266 |
| Consumer Welfare | 130 | 76 | 47 | 13 | 266 |
| Inequality Measures | 48 | 137 | 51 | 6 | 242 |
| Worker Satisfaction | 101 | 81 | 25 | 13 | 220 |
| Error Rate | 84 | 110 | 11 | 5 | 210 |
| Wages & Compensation | 98 | 47 | 30 | 10 | 185 |
| Regulatory Compliance | 88 | 73 | 17 | 7 | 185 |
| Automation Exposure | 66 | 64 | 33 | 16 | 182 |
| Team Performance | 105 | 29 | 30 | 11 | 176 |
| Training Effectiveness | 109 | 22 | 14 | 21 | 168 |
| Developer Productivity | 114 | 21 | 14 | 8 | 158 |
| Job Displacement | 12 | 90 | 24 | 1 | 127 |
| Hiring & Recruitment | 57 | 9 | 9 | 5 | 80 |
| Skill Obsolescence | 6 | 56 | 9 | 1 | 72 |
| Social Protection | 43 | 17 | 8 | 2 | 70 |
| Creative Output | 35 | 21 | 9 | 4 | 70 |
| Labor Share of Income | 18 | 21 | 17 | 1 | 57 |
| Worker Turnover | 15 | 16 | — | 4 | 35 |
| Industry | — | — | — | 1 | 1 |
Within the public sector, there is an emerging policy trend to incorporate AI considerations into workforce planning, including examining whether human positions may be substituted by technological solutions prior to recruiting new employees.
Paper reports an observed policy trend in public-sector workforce planning; specific policy documents, jurisdictions, or counts not provided in the excerpt.
Objectives, constraints, and prompt guidance affect reliability and generalization.
Authors' analysis and discussion based on experiments and ablations described in the paper (qualitative/empirical observations about sensitivity to objectives, constraints, and prompts).
The architect's role is shifting, but the human remains central.
Authors' discussion and interpretive analysis about the role of humans in agentic AI-driven design processes.
Across evolved designs, components often correspond to known techniques; the novelty lies in how they are coordinated.
Authors' qualitative analysis of evolved architectures and components reported in the paper (design inspection and interpretation of evolved solutions).
The study establishes statistically significant relationships between organizational AI adoption and compensation dynamics.
Econometric estimates (difference-in-differences and propensity score matched comparisons) using the combined datasets listed in the paper and controlling for industry, firm size, geography, occupation characteristics, and macroeconomic variables.
The study establishes statistically significant relationships between organizational AI adoption and changes in occupational structures.
Same econometric approach (difference-in-differences and propensity score matching) applied to combined datasets (Anthropic Economic Index, Census Business Trends and Outlook Survey, Federal Reserve regional surveys, labor market analytics), with controls for industry, firm size, location, occupation-level characteristics, and macroeconomic environment.
The study establishes statistically significant relationships between organizational AI adoption and changes in employment patterns in the United States during 2022–2025.
Econometric analysis using multiple large-scale data sources (Anthropic Economic Index, U.S. Census Bureau Business Trends and Outlook Survey, Federal Reserve regional surveys, labor market analytics) and methods described as difference-in-differences estimation and propensity score matching controlling for industry (NAICS 2-digit), firm size, geography, occupation characteristics, and macro conditions.
We identify significant differences between human and AI negotiation behaviors, finding that humans favor lower-complexity deals and are significantly less reliable partners compared to LM-based agents.
Results from the user study comparing human vs LM-based agent negotiation behavior (statements in the results section).
High-value uses require broader authority exposure — data access, workflow integration, and delegated authority — when governance controls have not yet decoupled capability from authority exposure.
Conceptual/mechanism claim articulated in the paper (motivating assumption for the analytical model; no empirical sample given in the abstract).
Firms are deploying more capable AI systems, but organizational controls often have not kept pace.
Stated as background context in the paper's abstract/introduction (observational claim; no empirical sample or experiment reported in the abstract).
The distribution of complementary (non-AI) skills across the workforce shapes whether AI improvements generate productivity bottlenecks or concentration-driven inequality.
Derived from the task-based model analysis described in the article; framed as a theoretical mechanism with reference to empirical patterns but without specific empirical study details in the excerpt.
There is a strict policy reversal in optimal editorial policy sign: tightening is optimal pre-transition, loosening is optimal post-transition.
Analytical proof in the model showing the sign reversal of the editor's optimal constrained response as AI capability crosses the critical threshold.
After the AI transition, editors must loosen acceptance standards while investing in AI detection, because further tightening only amplifies dissipative polishing without improving sorting.
Analytical characterization of the constrained optimal editorial response in the post-transition regime within the model; argument relies on the discontinuous reviewer-effort collapse and comparative statics.
The reviewer-effort collapse creates a welfare misalignment: authors benefit from a weakened 'rat race' while editors suffer from degraded signal informativeness.
Comparative statics and welfare analysis in the theoretical model showing authors' equilibrium payoffs rise as competition/polishing dissipates, while editor's signal informativeness declines due to lower reviewer effort.
In academic peer review, generative AI enters both sides of the market: authors use AI to polish submissions, and reviewers use it to generate plausible reports without exerting evaluative effort.
Model assumption and motivation in the paper's three-sided equilibrium framework; described as the dual adoption mechanism analyzed analytically (no empirical sample size reported).
The paper extends paradox theory to conceptualise the Creativity Paradox in the context of GenAI.
Theoretical extension and conceptual development within the paper (no empirical tests reported).
Within that n=11 subset, 9 of 11 agents shift by at least 2 ranks between composite and benchmark-only rankings.
Comparison of rank positions between composite and benchmark-only rankings on the 11-agent subset; reported count of agents that moved at least 2 ranks.
The four factors capture largely complementary information (n=50; ρ_max = 0.61 for Adoption-Ecosystem, all others |ρ| ≤ 0.37).
Correlation analysis among the four factor scores computed on the 50-agent sample; reported maximum inter-factor Pearson/Spearman correlation coefficients.
The intervention only modestly narrows the gap to a full-information benchmark.
Comparison between post-intervention calibration/auction outcomes and a full-information benchmark reported in the paper, showing only modest improvement.
Provisioned Throughput delivers the lowest latency at low concurrency but saturates its reserved capacity above approximately 20 concurrent users.
Empirical measurements from the instrumented system across concurrency up to 50 users and tier comparisons; the paper reports the observed saturation point near ~20 concurrent users.
Delegating tasks to genAI can be individually beneficial in the short term even as widespread adoption degrades future model performance (creating a social dilemma).
Result of the paper's behavioral model showing an individual-level incentive to use genAI versus a collective cost from adoption (theoretical/model-based; no empirical sample reported in abstract).
Token usage is highly variable and inherently stochastic: runs on the same task can differ by up to 30x in total tokens.
Observed run-to-run variability in total token counts for identical tasks across the collected agentic trajectories from eight frontier LLMs on SWE-bench Verified.
ASC (adaptive stopping criterion) halts harmful refinement but incurs a 3.8 pp confidence-elicitation cost.
Reported experiment with ASC showing that it prevents harmful iterative refinement yet causes a measured cost described as 3.8 percentage points due to confidence elicitation.
Only o3-mini (+3.4 pp, EIR = 0%), Claude Opus 4.6 (+0.6 pp, EIR ~ 0.2%), and o4-mini (+/-0 pp) remain non-degrading under self-correction; GPT-5 degrades by -1.8 pp.
Reported measured changes in accuracy (percentage-point changes) and measured EIR values for the named models after applying iterative self-correction across the experiment suite.
Across 7 models and 3 datasets (GSM8K, MATH, StrategyQA), we find a sharp near-zero EIR threshold (<= 0.5%) separating beneficial from harmful self-correction.
Empirical experiments reported across 7 LLMs and 3 benchmark datasets (GSM8K, MATH, StrategyQA) comparing outcomes of iterative self-correction as a function of measured EIR.
Firms with a high market position tend to imitate the peer leader, whereas firms in middle and low market positions are more likely to follow the peer group.
Heterogeneity analysis / subgroup regressions in fixed-effects models on panel data of publicly listed Chinese firms (2012–2023), stratifying firms by market position (high, middle, low).
AI influences innovation performance in organizations.
Discussion and synthesis of studies and reports on AI adoption and innovation performance presented in the review.
AI adoption is producing organizational implications, including changes in project management practices.
Findings synthesized from conference papers, case studies and industry reports included in the review.
Automation, generative AI, and intelligent systems are reshaping task structures, leading to both job displacement risks and the creation of new AI-driven roles.
Synthesis of empirical studies, conference findings, and industry reports reporting both displacement risks and new role emergence (review paper).
AI is rapidly transforming the nature of work, the demand for skills, and the professional roles of Information Technology (IT) practitioners.
Stated as a synthesis result from a narrative review of recent empirical studies, conference findings, and industry reports (review paper).
Semiconductors are a representative case study for analyzing weaponized interdependence in advanced technology sectors.
Methodological claim in the paper: selection and focus on the semiconductor sector as illustrative of broader advanced-technology sector dynamics under export restraints and chokepoint activation.
Previous literature is based primarily on the short-term effectiveness of coercion; this paper shifts attention to the longer-term structural consequences of technological restraints.
Literature review and positioning in the paper contrasting prior studies' short-term focus with the paper's longer-term structural emphasis (methodological/literature-critique claim).
Over time, U.S.–China reaction–counterreaction interactions generate three structural transformations: supply-chain reconfiguration, substitution, and regulations reinforcing segmentation.
Synthesis from the paper's longitudinal/case-analysis of semiconductor-related export restraints and subsequent industry and regulatory responses (qualitative identification of three emergent structural outcomes).
Current instability in U.S.–China relations arises less from complete ideological divergence or failure of outright containment policy than from a structured reaction–counterreaction dynamic triggered by chokepoint activation.
Argument based on qualitative analysis of U.S. export restraints after the first Trump administration and application of the 'weaponized interdependence' framework to advanced-technology sectors (paper's theoretical argument and case discussion).
AIGC is reshaping the rights and obligations of platforms and workers.
Argument in the paper describing legal and practical impacts of AIGC on platform-worker relationships; based on doctrinal/legal analysis and discussion of platform practices rather than reported quantitative empirical data.
The study explores implications of algorithmic enterprises for competitive advantage, labour markets, and regulatory policy.
Declared scope of the paper in the abstract; exploration is conceptual and analytical rather than reporting empirical findings or quantified effects.
Survey evidence suggests public attitudes towards AI combine optimism with apprehension, and most respondents oppose granting AI systems final authority over hiring and dismissal decisions.
Review cites multiple public opinion and survey studies reporting mixed (optimistic and apprehensive) attitudes and opposition to AI final authority in employment decisions (survey evidence summarized).
There are important regional differences—especially in developing contexts—that necessitate context-specific approaches to improving women’s participation in AI-enabled work.
Observation reported in the review drawing on geographically diverse studies and policy analyses; the abstract does not quantify differences or report sample sizes for cross-region comparisons.
Social, cultural, and ethical considerations influence women’s engagement in AI-centric workplaces.
Claim made in the review, based on interdisciplinary literature that includes sociocultural analyses and ethical discussions; the abstract does not provide empirical effect estimates or sample sizes.
AI applications—ranging from recruitment algorithms to workplace automation—can either reinforce gender disparities or promote equitable employment outcomes.
Stated in the review based on collated findings from multiple studies and analyses that document both harms (e.g., biased recruitment algorithms) and potential benefits (e.g., tools designed to reduce bias); no single empirical study or pooled effect size provided in the abstract.
Artificial Intelligence (AI) is rapidly transforming workplaces across the globe, offering both novel opportunities and unique challenges for women in technology-driven industries.
Stated in the paper's introduction/abstract as a summary conclusion based on a narrative literature review of peer-reviewed studies, policy analyses, and preprint research; no specific sample size or primary empirical method reported in the abstract.
The study proposes a sectoral risk classification to better understand vulnerability patterns and workforce implications.
Paper reports development/proposal of a sectoral risk classification as a contribution (the classification itself and validation details are not described in the abstract).
The rapid integration of Artificial Intelligence (AI) across industries is fundamentally reshaping occupational structures and redefining employment dynamics.
Stated as an overall conclusion of the paper based on a systematic review of recent literature from major academic databases (details of included studies not provided in the abstract).
These efficiency gains are offset by a growing 'Efficiency-Legitimacy Paradox' (i.e., improvements in efficiency come with worsening legitimacy concerns).
Conceptual synthesis from the systematic review (2018-2026) identifying a recurring trade-off across reviewed studies; specific empirical quantification not provided in abstract.
There is a structural shift from 'street level' bureaucracies to 'system-level' architectures that can be defined as the institutional division of 'Artificial Discretion' to algorithmic infrastructures.
Synthesis from the PRISMA-guided systematic review of literature (2018-2026) reporting observed changes in administrative architectures; specific studies not enumerated in abstract.
As a General-Purpose Technology (GPT), Artificial Intelligence (AI) is fundamentally reconfiguring state capacity, as well as the mechanics of global economic management.
Systematic review of current research studies (2018-2026) conducted following PRISMA guidelines; synthesis of literature claiming broad institutional and macroeconomic effects. Number of studies not specified in abstract.
Agentic AI differs from traditional algorithmic trading and generative AI through its capacity for goal-oriented autonomy, continuous learning, and multi-agent coordination.
Analytic comparison and synthesis across prior research and technical architectures in the survey; descriptive/definitional rather than empirical testing.
Uncertainty-aware exploration (in algorithms) alters fairness metrics compared to policies that ignore uncertainty.
Results from simulation experiments compare uncertainty-aware exploration policies to baseline policies and report changes in fairness metrics (as described in the abstract and results).
Analysis of more than two decades of M&A deals reveals shifts in acquisition activity and allows mapping of corporate linkages and overlapping investments.
Empirical longitudinal analysis of M&A deals over a period exceeding 20 years; method: mapping corporate linkages from M&A data (sample size/dataset not specified in the excerpt).
The emissions effects of digital trade are conditional rather than uniform, depending on complementary policy (carbon pricing, regulatory stringency), technological (AI-enhanced logistics), and energy (renewables) factors.
Synthesis of findings from fixed-effects regressions with interactions, carbon-pricing threshold analysis, machine-learning threshold detection, and SEM mediation on the monthly panel of 38 OECD economies (2000–2024).