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 |
Automation holds significant potential for modernising tax administration, but its success depends on aligning technological innovation with inclusive policy design and institutional capacity.
Overall conclusion of the literature synthesis of 36 peer-reviewed articles; based on patterns of positive impacts conditional on contextual factors and governance highlighted across the studies.
Behavioural responses to automation vary across taxpayer segments: some users embrace automation as a facilitator of compliance while others resist due to perceived opacity and technological anxiety.
Synthesis of behavioural findings from the reviewed literature (36 studies) reporting heterogeneous responses by taxpayer segment, including qualitative reports of resistance and quantitative measures of uptake/adoption.
The effectiveness of automated tax systems is mediated by contingencies including digital literacy, institutional trust, and regulatory clarity.
The review identifies recurring contextual factors across the 36 articles that are reported to moderate or mediate the impact of automation on outcomes (qualitative and quantitative findings cited in the synthesis).
The study identifies the main AI-enabled mechanisms advancing CE principles in smart manufacturing, waste valorisation, supply-chain transparency, and sustainable design.
Bibliometric network analysis of 196 peer-reviewed articles (2023–2024) and systematic review of 104 studies, per the abstract; identification is presented as a product of these analyses.
AI is not an inherent instrument of justice but a malleable socio-technical force whose equitable outcomes depend on policy design and institutional context.
Interpretation and synthesis of empirical results showing conditional and heterogeneous effects of AI; normative conclusion drawn by authors from observed heterogeneity and mediating channels.
Governmental structures, labor supply and demand, and incorporation of financial measures act as key intervening variables affecting achieved ROI from GenAI implementations.
Qualitative synthesis and theoretical analysis reported in the paper identifying contextual/intervening variables.
There is an evident tension between privacy and security in existing AI governance approaches.
Thematic synthesis and co-occurrence network from the reviewed studies identify trade-offs and tensions reported between privacy-preserving approaches and security requirements.
Generative AI serves as an effective 'wingman' for employment lawyers, capable of replacing substantial junior associate work while requiring continued human expertise for client counseling, supervision, and final legal advice preparation.
Authors' synthesis of experimental results showing AI-produced substantive analysis plus discussion about remaining limitations (e.g., citation errors) and required human oversight; qualitative assertion about substitutability for junior associate tasks.
The fragility of 'Pax Silica' has implications for global capitalism, technological governance, and geopolitical stability.
Analytical inference and concluding assessment based on theoretical framework and comparative analysis; no empirical quantification provided in the abstract.
The paper proposes new mechanisms through which big data affects individual welfare (beyond simple productivity gains), linking privacy costs, multiplier effects, and R&D transformation patterns.
Theoretical/mechanism development: the paper articulates new channels in its macro theoretical framework describing how data sharing impacts welfare via multiple mechanisms (model construction and analytic discussion; no empirical/sample validation).
Consumption is affected by the multiplier effect and the transformation patterns of R&D.
Theoretical: model analysis links consumption dynamics to a multiplier effect and to how R&D transforms inputs/outputs (comparative statics/dynamics in the theoretical framework).
Individuals’ welfare is influenced by both the privacy cost of big data sharing and their consumption levels.
Theoretical: welfare in the model is specified as a function of consumption and a privacy cost term arising from big data sharing; result follows from analytic derivation within the model (no empirical/sample data).
PPS gains are task-dependent: gains are large in high-ambiguity business analysis tasks but reverse in low-ambiguity travel planning tasks.
Task-level analysis across the three domains (business, technical, travel) within the controlled study (60 tasks total); authors report differential performance patterns by domain/ambiguity.
Capability and trust formally diverge beyond a critical scale (Capability-Trust Divergence).
Claim of a formal proof in the paper (mathematical / theoretical demonstration). No empirical sample size reported in the excerpt.
The Institutional Scaling Law shows that institutional fitness -- jointly measuring capability, trust, affordability, and sovereignty -- is non-monotonic in model scale, with an environment-dependent optimum N*(ε).
Theoretical derivation / analytic model presented in the paper (formal derivation of an 'Institutional Scaling Law'). No empirical sample size reported in the excerpt.
AI usage has dual effects on employees: it can both enhance innovative behavior and predict disengagement, as revealed by a dual-path (SOR-based) model.
Interpretation/synthesis from the four-stage longitudinal study of 285 finance professionals using a dual-path model based on SOR theory (combining the mediation and moderation results).
We evaluate 14 LLMs under zero-shot prompting and retrieval-augmented settings and witness a clear performance gap.
Experimental evaluation reported in the paper: authors state they ran experiments on 14 different large language models, under zero-shot and retrieval-augmented configurations, and observed differing performance across models.
Artificial intelligence embedded in human decision-making can either enhance human reasoning or induce excessive cognitive dependence.
Stated as a conceptual claim in the paper's introduction/abstract; supported by the paper's conceptual framing (theoretical argument), no empirical sample or experimental data reported here.
Policy implication: smarter, better-coordinated green governance is needed to address the negative local impacts and the crowding-out interaction between AI and environmental regulation.
Policy recommendation drawn in the abstract based on the empirical spatial findings (negative local effects and negative interaction).
Substantial regional gaps persist: leading eastern provinces approach a UCEE value of 1.0 while some northeastern provinces remain below 0.1.
Regional UCEE index estimates from the Super-SBM model across the 30 provinces reported in the abstract.
The systemic implications of AI in finance depend less on model intelligence alone than on how agent architectures are distributed, coupled, and governed across institutions.
Central argumentative claim supported by the AFMM conceptual model and an illustrative empirical application described in the paper (modeling + event-study approach); no full-sample details provided in the excerpt.
The Agentic Financial Market Model (AFMM), a stylised agent-based representation, links agent design parameters (autonomy depth, heterogeneity, execution coupling, infrastructure concentration, supervisory observability) to market-level outcomes including efficiency, liquidity resilience, volatility, and systemic risk.
Presentation of a stylised agent-based model (AFMM) in the paper; conceptual modelling linking specified agent parameters to macro/market outcomes. No empirical sample size reported in the excerpt.
Financial AI agents can be described by a four-layer architecture covering data perception, reasoning engines, strategy generation, and execution with control.
Conceptual framework proposed by the authors (theoretical/architectural proposal); no empirical testing or sample size provided.
These productivity gains are most pronounced for lower-skilled workers, producing a pattern the authors call “skill compression.”
Cross-study pattern reported in the literature review: comparative evidence across worker-skill strata in multiple empirical papers showing larger relative gains for lower-skilled/junior workers; specific underlying studies and sample sizes are not enumerated in the brief.
Financial well-being is not an automatic byproduct of automated credit efficiency but an emergent outcome of architectural alignment among technology, borrower capability, and governance structures.
Theoretical conclusion drawn from empirical results showing mixed effects (positive on repayment and resilience, negative on stress) and significant moderation by human capability and institutional design.
Study 1 quantifies confirmation bias through controlled experiments on 250 CVE vulnerability/patch pairs evaluated across four state-of-the-art models under five framing conditions for the review prompt.
Controlled experiment described in the paper: 250 CVE vulnerability/patch pairs evaluated across four state-of-the-art LLMs under five prompt framing conditions.
Lightweight safeguards can reduce risk in some settings but do not consistently prevent severe failures.
Analysis of simulated interventions/safeguards within governance simulations showing reductions in certain risk metrics in some scenarios, but persistence of severe failures in others; assessment based on rubric-judged transcript segments.
There are large differences in corruption-related outcomes across governance regimes and specific model–governance pairings.
Observed heterogeneity in outcomes across different authority structures and model–governance pairings within the multi-agent simulations, evaluated via rubric-based scoring over 28,112 transcript segments.
These findings challenge the narrative of complete automation by AI and underscore the enduring importance of human expertise in data science.
Interpretation based on competition results where AI-only baselines underperformed relative to many participant teams and top solutions used human-AI collaboration.
These findings indicate a misalignment between the perceived benefit of AI writing and an implicit, consistent effect on the semantics of human writing, with potential implications for cultural and scientific institutions.
Synthesis and interpretation of the paper's empirical results (user study, essay revision experiments, and peer-review analysis); presented as the paper's broader conclusion.
Socioeconomic regression analysis confirms strong correlations between neighborhood racial composition and detection likelihood: Pearson r = 0.83 for percent White and r = -0.81 for percent Black.
Reported Pearson correlation coefficients from regression analysis between neighborhood racial composition variables and detection likelihood in the simulations.
A Conditional Tabular GAN (CTGAN) debiasing approach partially redistributes detection rates but cannot eliminate structural disparity without accompanying policy intervention.
Experimental comparison between baseline simulations and CTGAN-debiased synthetic data showing partial redistribution of detection rates; paper asserts remaining structural disparities.
Regional analysis shows inland regions remain capital-dependent, with an estimated (capital) elasticity of approximately 0.43.
Regional decomposition/estimation reported in the study comparing inland regions to coastal ones using the extended production function.
The paper formalizes the distinction using a signal-aggregation model in which an organization maintains an anchor belief and achieves agreement through two exclusion channels: (1) report shrinkage toward the anchor and (2) a tolerance rule that discards reports deviating beyond a threshold.
Analytical formal model presented in the paper specifying an anchor belief and two exclusion mechanisms; model assumptions and mechanisms are explicit in the theoretical development. No empirical sample.
Organizational cohesion is observationally ambiguous: it can arise either from genuine information integration (debate and synthesis of heterogeneous inputs) or from exclusionary processes (conformity pressure, gatekeeping, intolerance of dissent).
Conceptual argument and formal definition in the paper framing; supported by the analytic distinction introduced in the paper between integration and exclusion as alternative generative mechanisms for observed agreement. No empirical sample—argument is theoretical and illustrated by model construction.
The authors identify ten evaluation practices that teams use, ranging from lightweight interpretive checks to formal organizational processes (examples: qualitative user reviews, red-team testing, A/B experiments, telemetry/log analysis, structured annotation, governance/meta-evaluation).
Thematic coding of 19 interview transcripts produced a taxonomy enumerating ten practices (paper reports the taxonomy as an outcome).
Quantum-driven growth depends critically on adoption rates, infrastructure readiness, complementary investments (digital infrastructure, human capital), and enabling policy/regulatory environments.
Scenario framework that varies (a) technical timelines, (b) sectoral adoption rates (diffusion models), (c) infrastructure readiness, and (d) policy environments; policy counterfactual modeling shows sensitivity of adoption and macro outcomes to these parameters.
The magnitude and timing of macroeconomic impact from quantum computing are highly uncertain.
Monte Carlo / scenario ensemble results showing wide (fat-tailed) outcome distributions driven by uncertainty in technical milestones, adoption rates, and complementarity strengths; use of expert elicitation to parameterize tail risks.
Policymakers face trade-offs between promoting innovation and market efficiency on one hand and protecting privacy, fairness, and national security on the other; economic analysis can inform calibration.
Normative policy analysis and synthesis of literature on digital regulation and trade-offs; supported by comparative observations of regulatory priorities across jurisdictions.
Safeguards such as audit trails, explainability, and human oversight impose additional implementation costs that must be weighed against efficiency benefits.
Normative and economic reasoning based on requirements for compliance and system design; no empirical cost estimates provided.
There is a fundamental tension between AI-driven efficiency and core administrative-law principles—discretion, due process, and accountability.
Doctrinal legal analysis of administrative-law principles in Vietnam and comparative institutional analysis of AI adoption in other systems.
The net educational value of AI-generated feedback depends on alignment with pedagogical goals, quality evaluation, integration with human teaching, and governance to manage equity, privacy, and incentives.
Synthesis statement from the meeting report produced by 50 interdisciplinary scholars; conceptual judgment rather than empirical proof.
Convergence after exemplar exposure occurred by both tightening of estimates within a measure family and by agents switching measure families.
Agent-level tracking across stages showed two patterns following exemplar exposure: (1) reduced within-family dispersion (tighter estimates) and (2) categorical switches in measure selection by some agents, as recorded across the 150-agent sample.
LLMs excel at extracting and generating arguments from unstructured text but are opaque and hard to evaluate or trust.
Synthesis of recent LLM literature and observed properties (generation capability vs. opacity); no empirical evaluation within this paper.
The paper is primarily theoretical and historical; empirical validation is needed to quantify the irreducible component of LLM value, and practical degrees of rule‑extractability may exist even if some capabilities remain tacit.
Stated limitations section acknowledging the theoretical nature of the work and the need for empirical follow‑up.
If an LLM's full capability were reducible to an explicit rule set, that rule set would be an expert system; because expert systems are empirically and historically weaker than LLMs, this leads to a contradiction (supporting non‑rule‑encodability).
Logical proof‑by‑contradiction presented in the paper, supported by conceptual mapping between rule sets and expert systems and qualitative historical comparisons.
HindSight has limitations: it depends on citation and venue proxies for impact, uses a finite forward window (30 months), and may undercount delayed-impact research and be domain-specific to AI/ML.
Authors' stated limitations in the paper noting reliance on observable downstream signals (citations/venues), the finite forward window, field heterogeneity, and measurement noise.
Practical caveats: benefits depend on accelerators supporting MXFP formats; despite up to 96% recovery, residual quality gaps may remain for some task-specific or safety-critical cases; integration and tuning cost is required to apply BATQuant.
Discussion/limitation section in the paper outlining hardware dependency, remaining quality gaps despite high recovery percentages, and engineering effort for integration and tuning; these are argumentative caveats rather than results of controlled experiments.
The sign of the Largest Lyapunov Exponent (LLE) gives a precise criterion: negative LLE (contracting dynamics) permits fast convergence and real speedups for parallel Newton methods, whereas positive LLE (expanding/chaotic dynamics) prevents generally achieving fast convergence.
Theoretical derivation relating Lyapunov exponents to the stability of parallel-in-time linearizations and convergence of the parallel Newton iterations; supported by empirical observations reported on representative tasks.
Many fixed-point and iterative schemes (e.g., Picard, Jacobi) are unified as special cases within the parallel Newton framework.
Theoretical analysis and derivations in the thesis that show these classical iterative methods arise from particular choices/approximations in the parallel Newton formulation.