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
Human-AI collaboration task complexity (HAI-C task complexity) negatively affects employees' work engagement by amplifying their HAI-C tech-learning anxiety.
Three-wave longitudinal survey of matched data from 497 employees; mediation analysis using hierarchical regression and bootstrapping reported in the Results section.
LLMs are not only less accurate on ideologically contested economic questions, but systematically less reliable in one ideological direction than the other, underscoring the need for direction-aware evaluation in high-stakes economic and policy settings.
Synthesis of empirical findings: lower accuracy on contested items, higher accuracy for intervention-aligned cases in 18/20 models, and error skew toward intervention-oriented predictions; policy recommendation follows from these empirical patterns.
This directional skew is not eliminated by one-shot in-context prompting.
Intervention of one-shot in-context prompting applied to models; evaluation shows the intervention-oriented error skew persists despite one-shot prompting.
Ideology-contested items are consistently harder than non-contested ones.
Comparison of model performance (accuracy) on contested subset (1,056 items) versus non-contested items in the 10,490-triplet benchmark; reported consistent lower accuracy on contested items.
Important boundary conditions include data maturity, process integration, governance discipline, and the degree of functional trust between finance and operating units.
List of boundary conditions reported in the paper based on documentary case analysis and synthesis with literature.
GenAI does not improve management accounting decision quality primarily by replacing managerial judgment.
Interpretive finding based on documentary analysis of disclosures from the three case firms and relevant literature; presented as a summary conclusion in the paper.
The stakes are particularly high in spreadsheet environments, where process and artifact are inseparable: each decision the agent makes is recorded directly in cells that belong to and reflect on the user.
Conceptual / domain-specific argument made by the authors (no empirical sample attached to the claim).
AI agents can perform sophisticated, multi-step knowledge work autonomously from start to finish, yet this process remains effectively inaccessible during execution: by the time users receive the output, all underlying decisions have already been made without their involvement.
Author assertion / conceptual description in the paper (no empirical quantification provided for this general statement).
Advances in AI agent capabilities have outpaced users' ability to meaningfully oversee their execution.
Author assertion / literature-level observation presented in the paper (no empirical sample reported for this claim).
Selective forgetting remains underexplored compared to retention in LLM agent memory research.
Authors' literature survey / position statement in paper (assertion made in abstract).
Beyond technical barriers there are organizational ones: a persistent AI literacy gap, cultural heterogeneity, and governance structures that have not yet caught up with agentic capabilities.
Interview data (over 30) reporting organizational challenges including limited AI literacy, diverse cultural attitudes across organizations, and lagging governance relative to agentic AI capabilities.
Adoption is constrained less by model capability than by fragmented and machine-unfriendly data, stringent security and regulatory requirements, and limited API-accessible legacy toolchains.
Stakeholder interviews (over 30) reporting barriers to deployment; qualitative synthesis identifies data fragmentation, security/regulatory requirements, and legacy toolchain access as primary constraints.
Providing agents feedback about past performance makes them worse at information aggregation and reduces their profits.
Experimental condition where agents received feedback about past performance; compared aggregation (log error of last price) and profits with and without feedback and found worse aggregation and lower profits when feedback was given.
Increasing the complexity of the information structure has a significant and negative impact on information aggregation, suggesting AI agents may suffer from the same limitations as humans when reasoning about others.
Experimental manipulation of information-structure complexity in the controlled trading experiment; measured change in aggregation performance (log error of last price) as complexity increases.
Users push back against agent outputs -- through corrections, failure reports, and interruptions -- in 44% of all turns.
Turn-level coding of user behavior in the SWE-chat dataset: proportion of conversational turns containing correction/complaint/interrupt signals, computed across >63,000 user prompts and sessions.
Agent-written code introduces more security vulnerabilities than code authored by humans.
Comparative analysis of security vulnerabilities attributed to agent-authored code versus human-authored code within the SWE-chat dataset (method details not specified in excerpt).
Just 44% of all agent-produced code survives into user commits.
Empirical measurement of code provenance and survival within the SWE-chat dataset: proportion of agent-produced code that becomes part of subsequent user commits across sessions.
Despite rapidly improving capabilities, coding agents remain inefficient in natural settings.
Authors' summary claim supported by dataset-derived metrics such as agent code survival rate (44%) and user pushback (44% of turns); observational analysis of SWE-chat.
Regulated deployment imposes four load-bearing systems properties — deterministic replay, auditable rationale, multi-tenant isolation, statelessness for horizontal scale — and stateful architectures violate them by construction.
Conceptual/architectural argument presented in the paper (theoretical analysis), not an empirical measurement in the abstract.
Evaluation of four leading AI platforms shows that standard RAG-based approaches achieve an average of only 15% accuracy when information is insufficient.
Empirical evaluation described in paper: four AI platforms tested on benchmark; reported average accuracy of 15% for RAG-based approaches on cases with insufficient information.
Unemployment insurance adjudication has seen rapid integration of AI systems and the question of additional fact-finding poses the most significant bottleneck for a system that affects millions of applicants annually.
Contextual/introductory claim in paper; references to domain-scale impact and bottleneck; no specific numeric study sample provided in excerpt.
A well-known limitation of AI systems is presumptuousness: the tendency of AI systems to provide confident answers when information may be lacking.
Statement in paper framing the problem; general literature/contextual claim (no specific experiment cited in the excerpt).
Brevity, semantic isolation and rhetorical register independently predict representational outcome (i.e., which submissions are included/excluded in summaries).
Statistical/semantic analysis (presumably regression or causal inference) reported in the paper linking textual features—brevity, semantic isolation, rhetorical register—to representational outcomes.
Exclusion concentrates in clusters expressing dissent, scepticism and critique of AI, with exclusion rates of 33%–88% in such clusters.
Cluster/semantic analysis reported in the paper showing higher exclusion rates for clusters labeled as dissent/scepticism/critique.
In topic B, 15.3% of participants are effectively excluded by the official summary.
Empirical measurement reported in the paper quantifying participants 'effectively excluded' when comparing source submissions to official summary coverage.
In topic A, 16.9% of participants are effectively excluded by the official summary.
Empirical measurement reported in the paper quantifying participants 'effectively excluded' when comparing source submissions to official summary coverage.
Both official government summaries underperform a random-participant baseline for topic B (coverage degradation of -8.0%).
Empirical comparison in the paper between official government summary and a random-participant baseline using the n=5,253 consultation responses.
Both official government summaries underperform a random-participant baseline for topic A (coverage degradation of -9.1%).
Empirical comparison in the paper between official government summary and a random-participant baseline using the n=5,253 consultation responses.
LLMs endorsed fraudulent investments at 0% across all models tested.
Preregistered experiment across seven leading LLMs producing 3,360 AI advisory conversations; reported 0% endorsement of objectively fraudulent opportunities.
Endorsement reversal occurred in fewer than 3 in 1,000 observations.
Observed incidence reported from the preregistered experiment (3,360 AI advisory conversations); statement in paper reporting incidence <3/1,000.
The policy and research challenge posed by platform-mediated automation is not merely job quantity (technological unemployment) but institutional continuity — how societies reproduce practical competence when platforms optimize for efficiency rather than formation.
Normative and conceptual claim developed through literature synthesis (institutional economics, platform governance, workforce development); presented as an analytical reframing rather than an empirically tested hypothesis.
Entry-level roles have historically functioned as apprenticeships in which workers acquire tacit knowledge and critical judgment; if platforms curtail these formative occupational layers, organizations may lack future workers capable of exercising contextual reasoning required to manage complex systems.
Institutional economics and workforce development literature cited in the paper; conceptual synthesis without original empirical measurement reported.
Platform-mediated automation risks hollowing out labor structures from both directions: eroding repetitive, junior roles from below and automating supervisory coordination functions from above.
Theoretical argument synthesizing institutional economics and platform literature; articulated as a conceptual risk rather than demonstrated with original empirical data.
Algorithmic systems are displacing routine tasks across both low-wage entry-level work and middle-management functions.
Stated in paper's argumentation; supported by a literature-based review drawing on platform governance literature and recent research on AI-enhanced automation (no original empirical sample or quantitative study reported).
A gender gap persists, concentrated in the most exposed occupations.
Stratified/descriptive and regression analyses of the 2024 EWCS showing gender differences in self-reported generative AI adoption, with the gap largest among occupations with highest exposure; sample >36,600 workers across 35 countries.
The infrastructure for cross-user agent collaboration is entirely absent, let alone the governance mechanisms needed to secure it.
Authoritative claim in paper framing the research gap; presented as observational/argumentative (no empirical audit reported).
Current AI agent frameworks have made remarkable progress in automating individual tasks, yet all existing systems serve a single user.
Statement in paper's introduction/positioning; conceptual survey-style claim (no empirical study or systematic benchmark reported).
Standard benchmarks often fail to isolate an agent's core ability to parse queries and orchestrate computations.
Paper asserts that existing/standard benchmarks do not adequately isolate parsing and computation-orchestration abilities, motivating the new benchmark.
As multimodal AI achieves human-parity understanding of speech and gesture, [the keyboard's] necessity dissolves.
Theoretical claim supported by multidisciplinary review (history, neuroscience, technology, organizational studies); no quantified empirical test reported.
General-purpose LLMs pose misinformation risks for development and policy experts, lacking epistemic humility for verifiable outputs.
Conceptual/argumentative claim stated in the paper's motivation; no empirical test reported in the abstract.
There was a nonsignificant absolute retest performance reduction in the AI condition and a larger retest performance decrement in the AI condition (i.e., retention decreased more after using Copilot).
Comparison of retest (one-week) performance across conditions reported in results; authors report a nonsignificant reduction and larger decrement for the AI/Copilot condition (n=22).
Current operational approaches typically involve scattered testing tools, resulting in partial coverage and errors that surface only after deployment.
Authors' characterization of industry practice and limitations (assertion in paper; no empirical sample size reported in abstract).
Network change validation remains a critical yet predominantly manual, time-consuming, and error-prone process in modern network operations.
Statement in paper framing the problem; based on authors' characterization of current operational practice (no empirical sample size reported in abstract).
Thick subjectivist theories of meaning in life and meaningful work—those theories that emphasize that meaning-conferring activities are historically formed—enable us to appreciate how some losses cannot be made up, even if there are in principle ample alternative sources of meaning to be found elsewhere.
Theoretical claim about the explanatory power of 'thick subjectivist' normative theories; argued via conceptual philosophical analysis in the paper (no empirical testing reported).
Even if there are rich non-work sources of meaning, this does not entail that there is not a significant and multi-faceted loss of meaning, one that cannot be compensated for or offset elsewhere.
Normative/philosophical argument presented in the paper (conceptual reasoning rather than empirical measurement; no sample size).
The argument that non-work goods can replace work-derived meaning fails to consider the embeddedness and thickness of meaning in human lives.
Philosophical/theoretical critique based on conceptual analysis (author's argument invoking the notions of embeddedness and thickness of meaning; no empirical study reported).
The paper identifies governance challenges such as accountability gaps, digital sovereignty risks, ethical pluralism, and strategic weaponization arising from embedding AI in diplomatic practice.
Conceptual and normative analysis section of the paper outlining risks and governance challenges; illustrated by examples and argumentation.
Thin training coverage fosters anxiety about substitution and slows diffusion of AI tools.
Reported associations from surveys of mid-level managers and technical staff, interviews, and document analysis across cases; thematic coding identified links between limited training, worker anxiety, and slower diffusion. (Sample size not reported.)
Upstream textile SMEs frequently exhibit constrained supply chain resilience owing to persistent information latency and structural dependence on downstream orders.
Background/contextual claim stated in paper (motivation for study); no specific quantitative test reported in abstract.
Platforms can exploit workers' uncertainty about the cost of labor to effectively suppress wages.
Interpretation / implication drawn from the theoretical model and the result that a platform can achieve coverage while paying only O(log(M)/M) fraction of total labor cost under assumptions about workers' cost estimates.