Evidence (7560 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
Filtered →
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 |
Human Ai Collab
Remove filter
Even a high-quality initial model can be undermined over time because human creators, relying on the model, produce more homogeneous content that harms subsequent training and lowers model performance.
Dynamic theoretical modeling and qualitative analysis in the paper demonstrating increased reliance on AI-assisted content creation and its effect on training data distribution.
In a dynamic model, an initially good AI model induces greater human reliance on AI-assisted creation, which homogenizes content and creates a feedback loop that degrades the model's own performance — a phenomenon termed the "curse of precision."
Dynamic theoretical model and analysis presented in the paper showing feedback effects across training rounds; no empirical sample reported.
More innovative creators are especially harmed under the strong-IP regime — a phenomenon the paper terms the "originality penalty."
Analytical result derived from the static game model in the paper highlighting differential effects by creator innovativeness; theoretical characterization labeled "originality penalty."
A regime of strong intellectual property rights, modeled as a static Stackelberg game, also fails to provide adequate creative incentives (it underpowers creative incentives).
Theoretical analysis using a static Stackelberg-game model developed in the paper; analytical results show reduced creator incentives under this regime.
A "free-for-all" model based on fair use fails because it does not compensate creators for their contributions.
Analytical / conceptual argument presented in the paper (no empirical sample); model-based reasoning showing creators receive no compensation under a free-for-all regime.
Unequal access to GenAI tools in higher education may exacerbate employability gaps and inequities among students.
Concern identified and discussed in the literature as summarized in the review article (conceptual/literature-based; no new empirical evidence reported).
GenAI raises concerns including passive dependence, weakened critical thinking, uncertain authorship, academic integrity breaches, algorithmic bias, unequal access, and employability gaps.
Synthesis of concerns reported in prior studies and discussions in the literature (review article); no new empirical data provided.
Agents do not reroute along the methodological axes humans use to bias their estimates.
Analytic comparison showing that when biased prompts are applied, agents change methodological choices but not along the particular methodological axes (choices) that human analysts exploit to bias estimates.
Benchmarking multiple state-of-the-art open and closed source VLMs on our evaluation framework demonstrates substantial limitations in current engineering reasoning capabilities.
Empirical claim based on the paper's benchmarking experiments using the EngVQA dataset and the 8-stage framework (models and detailed results not provided in the excerpt).
Existing benchmarks primarily evaluate final answers and provide limited assessment of intermediate reasoning processes.
Claim in paper contrasting EngVQA's process-oriented evaluation with prior benchmarks (literature/benchmark review claim; no specific benchmarks or quantitative comparison provided in the excerpt).
Failures in engineering reasoning by AI systems may produce physically invalid yet superficially plausible solutions, posing risks for engineering education, scientific assistance, and technical decision-making.
Argumentative claim in the paper highlighting potential risks of reasoning failures in high-stakes engineering contexts (motivational/background statement).
Datasets are rarely standardized or shared.
Review synthesis and commentary across included studies and supplementary documents indicating limited data standardization and sharing.
Penerapan AI menimbulkan isu etika dan keamanan data yang memerlukan tata kelola AI yang bertanggung jawab.
Sistematis studi literatur yang menelaah 33 sumber ilmiah, laporan lembaga internasional, dan kebijakan terkait (n=33).
AI meningkatkan risiko pengangguran pada sektor yang pekerjaannya bersifat rutin.
Sistematis studi literatur yang menelaah 33 sumber ilmiah, laporan lembaga internasional, dan kebijakan terkait (n=33).
Penerapan AI menyebabkan kesenjangan keterampilan (skill gap) antara kebutuhan pasar dan kemampuan tenaga kerja.
Sistematis studi literatur yang menelaah 33 sumber ilmiah, laporan lembaga internasional, dan kebijakan terkait (n=33).
Neither MCP nor A2A defines the shared workspace in which humans and agents perform accountable work together.
Analytical claim by the authors contrasting existing standards with the missing specification of a shared human-agent workspace; no empirical evaluation provided.
In current practice the human judgement is recorded, if at all, in application code, chat threads, ticket comments, and tribal memory.
Descriptive statement about current recording practices; presented without empirical study or counts in the provided text.
The technical surface for this collaboration remains weakly specified.
Asserted by the authors as an assessment of current technical standards and interfaces; no audit or measurement cited in the provided text.
Automated evaluators on which the community currently relies -- LLM-as-a-judge, artificial metrics, and even state-of-the-art (SOTA) models -- agree weakly with expert judgment.
Comparisons/correlations between automated evaluator outputs and human expert ratings showing weak agreement.
LLMs falter most in pluralistic fields like the social sciences that demand context-aware interpretation and evolving theories.
Field-level performance comparisons showing lower agreement/quality of LLM-generated ideas in pluralistic fields (social sciences) relative to more constrained fields (life sciences etc.).
Senior social scientists are the harshest critics, and their skepticism is well-earned.
Subgroup analysis by seniority and field indicating senior social scientists give lower ratings (more critical) than other subgroups; interpretation provided by authors.
Non-reasoning LLMs collapse into a narrow 'hivemind' of similar ideas.
Comparative analysis of idea outputs from different LLM classes showing reduced diversity/similarity concentration for non-reasoning models (as described in results).
This regulatory pressure creates a direct conflict between multi-stakeholder transparency and corporate data privacy.
Paper's conceptual argument describing a tension between transparency requirements and proprietary data protection; no empirical study provided.
Regulatory compliance demands have surpassed the capacity of manual corporate reporting.
Assertion in paper (conceptual observation about reporting capacity); no empirical measurement or sample size reported.
The convergence of the 2026 European Union Safe and Sustainable by Design (SSbD) framework, Corporate Sustainability Due Diligence Directive (CSDDD), and Carbon Border Adjustment Mechanism (CBAM) introduce a severe governance bottleneck for advanced semiconductor manufacturing facilities ("Smart Fabs").
Declarative claim in paper based on policy convergence analysis; no empirical dataset or sample size reported (conceptual/analytical argument).
Learning specialized simulator input languages can cost domain scientists hours to days.
Stated motivating claim in the paper (no experimental sample size or formal measurement reported in abstract).
Transformational leadership negatively moderates the relationship between AI application and employees' job insecurity, buffering employees' insecurity responses across varying levels of AI application.
Moderation analysis reported in the study using the same employee survey dataset (411 valid responses), indicating a statistically significant buffering (negative) moderating effect of transformational leadership on the AI–job insecurity relationship.
Self-efficacy negatively moderates the relationship between AI application and employees' job insecurity by strengthening the insecurity-reducing effect of moderate AI application and weakening the insecurity-enhancing effect of excessive application.
Moderation analysis on the same cross-sectional survey data (411 valid employee questionnaires), reporting a statistically significant negative (buffering) interaction of self-efficacy with AI application intensity on job insecurity.
In hyperscale cloud network infrastructure, traditional human-driven incident response cannot keep pace with the volume, velocity, and complexity of failures.
Stated as background/motivation in the paper; no quantitative data, sample size, or empirical comparison provided in the abstract.
Agents frequently overlook subtle yet critical details that are obvious to real human researchers.
Reported as a qualitative result/observation from the authors' experiments on AARRI-Bench; no numeric frequency or sample size provided in the excerpt.
Extensive experiments across frontier models and agentic systems reveal that even the best-performing configuration (Mini-SWE-Agent with Claude Opus 4.7) achieves only a 68.3% success rate on AARRI-Bench.
Empirical evaluation reported in the paper: experiments across multiple models/agentic systems; the excerpt reports the top configuration and its success rate. The excerpt does not state the number of tasks or sample size.
Despite their evolution from research assistants into autonomous research agents, these systems still exhibit significant limitations in field sensitivity, research ethics, and nuanced scientific judgment, and consequently remain unable to fully replace human researchers.
Asserted in the paper as a high-level observation and motivation; the excerpt does not provide quantified evidence or sample sizes for these limitations.
Existing approaches to explainability are predominantly post-hoc, offering unstable, non-contestable accounts that have no formal relationship to the reasoning process that produced the output.
Critical literature/argumentative claim in the paper; presented as a conceptual critique rather than supported by empirical evaluation or systematic review data.
Opacity of such models in these settings is not merely inconvenient but institutionally and legally untenable.
Normative/legal argument presented in the paper (conceptual reasoning about institutional and legal requirements); no empirical legal-case analysis or quantified legal rulings provided.
Employees currently lack clear guidance on appropriate use of GenAI within organizations.
Background claim in paper motivating the study (statement that employees 'lack clear guidance on appropriate use').
Current research on AI-supported conflict techniques has focused predominantly on Devil's Advocate (DA) and has neglected Dialectical Inquiry (DI).
Literature review / gap statement in the paper pointing to relative emphasis on DA in prior research and lack of work on DI.
Other methods, such as variants of prediction-powered inference, do not have the 'do no harm' guarantee.
Comparative methodological claim in the paper (abstract)—likely supported by theoretical discussion and comparisons in the main text.
Even a perfect non-proprietary-data report would be capped at 3.83 by B's coverage (i.e., B imposes an upper bound on non-proprietary informed decision-quality).
Analytic upper-bound calculation based on B's measured coverage on the curated gold record (exact derivation not provided in abstract).
Interviews provide expanded analysis on existing skill gaps and lifelong learning needs among wind-energy professionals.
Qualitative interview data are reported to highlight skill gaps and lifelong learning needs; specific counts of interviewees not provided in the summary.
Operators have no standard way to tell an autonomous agent that a resource is off-limits: access controls either let the agent in (it has valid credentials) or hard-fail it (indistinguishable from any other client).
Analytical description/argument presented in the paper (problem statement); no empirical data reported for this claim.
Reporting on ethics, transparency and governance was inconsistent.
Reported synthesis result from the scoping review noting variability in reporting practices across included studies.
Formal theory use was limited, with only a small minority of studies explicitly drawing on established frameworks.
Authors' assessment of methodological/theoretical characteristics of the included empirical studies in the scoping review.
Fewer studies evaluated individual-facing developmental support, and sustained career outcomes were rarely measured.
Reported gap identified in the scoping review findings summarised in the abstract.
Empirical evidence on applications designed to support women’s career development remains limited.
Conclusion drawn from the scoping review: authors searched seven databases + backward/forward citation searching and synthesised identified empirical studies.
The framework reframes the education–employer gap as a structural failure in the pathway and outlines implications for universities, employers, accreditors, and policymakers.
Conceptual claim and implications drawn by the author(s) in the paper (stated in the abstract).
The architecture of the undergraduate degree is structurally incapable of replacing the informal post-degree apprenticeship system through curricular revision alone.
Argument presented in the paper, supported by the systematic review of eighteen peer-reviewed studies and labor-market analyses cited in the abstract.
The informal post-degree apprenticeship system that historically completed graduate formation no longer reliably exists.
Claim based on the paper's systematic review of eighteen peer-reviewed studies and current labor-market analyses (as described in the abstract).
Higher education has misdiagnosed the resulting challenge as curriculum misalignment—a content problem assumed to be solvable through revised syllabi, AI electives, and marginal expansions of experiential learning.
Argument presented in the paper, supported by the paper's systematic review of eighteen peer-reviewed studies and labor-market analyses (as described in the abstract).
Artificial intelligence and automation are restructuring early-career knowledge-work roles by compressing the entry-level functions through which graduates historically built portfolios, developed professional judgment, and earned professional credibility.
Statement supported in the paper by a systematic review of eighteen peer-reviewed studies and current labor-market analyses (as described in the abstract).
GenAI usage significantly decreased creativity-relevant skills.
Experiment with 82 participants reported in the paper; authors report a statistically significant decrease in measures of creativity-relevant skills for participants using GenAI.