Evidence (7198 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
8921 claims
Filter claims →
Productivity
8002 claims
Filter claims →
Governance
7198 claims
Filtered →
Human-AI Collaboration
6864 claims
Filter claims →
Org Design
4398 claims
Filter claims →
Innovation
4286 claims
Filter claims →
Labor Markets
3629 claims
Filter claims →
Skills & Training
3001 claims
Filter claims →
Inequality
2141 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 | 790 | 208 | 103 | 950 | 2117 |
| Governance & Regulation | 869 | 411 | 195 | 126 | 1630 |
| Organizational Efficiency | 817 | 202 | 126 | 87 | 1243 |
| Technology Adoption Rate | 675 | 258 | 128 | 106 | 1178 |
| Research Productivity | 462 | 138 | 64 | 347 | 1023 |
| Output Quality | 501 | 193 | 61 | 52 | 807 |
| Decision Quality | 346 | 180 | 84 | 51 | 668 |
| AI Safety & Ethics | 235 | 285 | 70 | 34 | 630 |
| Firm Productivity | 452 | 58 | 91 | 20 | 627 |
| Market Structure | 184 | 171 | 123 | 24 | 507 |
| Task Allocation | 221 | 65 | 76 | 34 | 401 |
| Skill Acquisition | 176 | 62 | 62 | 17 | 317 |
| Innovation Output | 207 | 28 | 48 | 18 | 303 |
| Fiscal & Macroeconomic | 135 | 72 | 44 | 26 | 284 |
| Employment Level | 105 | 56 | 108 | 13 | 284 |
| Consumer Welfare | 121 | 67 | 45 | 11 | 244 |
| Firm Revenue | 160 | 50 | 28 | 4 | 242 |
| Task Completion Time | 182 | 33 | 10 | 13 | 239 |
| Inequality Measures | 45 | 126 | 50 | 6 | 227 |
| Worker Satisfaction | 94 | 73 | 23 | 12 | 202 |
| Error Rate | 76 | 98 | 11 | 4 | 189 |
| Regulatory Compliance | 81 | 73 | 17 | 7 | 178 |
| Automation Exposure | 61 | 59 | 26 | 14 | 163 |
| Training Effectiveness | 97 | 21 | 14 | 19 | 153 |
| Wages & Compensation | 78 | 37 | 25 | 6 | 146 |
| Developer Productivity | 105 | 18 | 14 | 6 | 144 |
| Team Performance | 87 | 17 | 28 | 10 | 143 |
| Job Displacement | 12 | 83 | 21 | 1 | 117 |
| Hiring & Recruitment | 52 | 8 | 8 | 3 | 71 |
| Social Protection | 39 | 17 | 8 | 2 | 66 |
| Creative Output | 32 | 20 | 8 | 3 | 64 |
| Skill Obsolescence | 5 | 49 | 6 | 1 | 61 |
| Labor Share of Income | 17 | 19 | 17 | — | 53 |
| Worker Turnover | 15 | 14 | — | 3 | 32 |
| Industry | — | — | — | 1 | 1 |
Governance
Remove filter
Policy frameworks, reskilling initiatives, and institutional adaptations are required to ensure inclusive technological progress.
Prescriptive conclusion presented in abstract based on the review and synthesis; no empirical validation or sample sizes provided in abstract.
AI simultaneously generates demand for higher-order problem solving, emotional intelligence, and human-AI collaboration skills.
Explicit finding reported in abstract from the review of interdisciplinary literature; no quantified effect sizes or sample sizes provided in abstract.
Latency relaxation expands feasible geography for placing inference workloads.
Result reported from the paper's modeling and stylized simulation (energy-latency frontier analysis showing marginal cost/carbon benefits from relaxing latency budgets).
The paper provides a transparent stylized simulation over representative global compute regions to show how heterogeneous latency tolerance separates workloads into local, regional, and energy-oriented execution layers.
Empirical/methodological evidence from a stylized simulation described in the paper; uses representative global compute regions and latency-tolerance heterogeneity to categorize workloads.
AI inference is becoming a persistent and geographically distributed source of electricity demand.
Statement/assertion in the paper's introduction framing the motivation; no empirical sample or experiment reported in the provided text.
Effective governance requires coordinated action across technical, organizational, and regulatory domains (e.g., system-level audits, vendor guidelines, continuous monitoring, documentation across dependency chains) to establish meaningful accountability in distributed development environments.
Policy and technical recommendations derived from literature review, regulatory analysis, and the paper's conceptual findings (recommendation, not empirically validated).
Managing evolutionary dynamics in software is as urgent as AGI alignment for safeguarding society’s co-evolution with its machines.
Author's concluding normative claim in the abstract; argument based on scenario analysis rather than comparative empirical evidence.
Governance should shift focus from aligning goals to steering evolution; the paper proposes four guidance instruments: replication-rate thresholds (modeled on epidemiological R0), a public vulnerability registry for self-modifying code, tiered digital biosafety levels, and adaptive regulatory sandboxes.
Normative policy recommendation spelled out in the abstract; based on the paper's scenario analysis and argumentation rather than empirical validation.
Cloud platforms, open-source software supply chains, and crypto-economic incentives provide, at electronic speed, the three preconditions of evolution: replication, variation, and differential fitness.
Conceptual/mechanistic claim supported by theoretical argumentation and scenario-building in the paper (no empirical test or sample reported).
The proposed framework balances AI-driven productivity with the epistemic sovereignty necessary to manage increasingly opaque software ecosystems.
Normative/architectural claim about the proposed framework; presented conceptually in the paper without reported empirical testing in the excerpt.
To preserve long-term resilience, engineering leaders must move beyond prompt-based development to implement rigorous human-in-the-loop pedagogical standards.
Prescriptive recommendation based on the paper's conceptual analysis; no randomized trials or empirical validation of this intervention reported in the excerpt.
Comparing the verbal-profile setting to a numeric-budget condition with confidentiality instructions cleanly isolates role coherence as distinct from instruction-following failure.
Experimental comparison between verbal-profile condition and numeric-budget condition with confidentiality instructions; result claimed to isolate mechanism (role coherence) from mere instruction-following failure.
In an experiment where a language-model buyer agent shops on behalf of a verbal consumer profile, seller-side inference from dialogue alone recovers willingness to pay nearly one-for-one.
Reported experimental result using a language-model buyer agent interacting on behalf of a verbal consumer profile; experimental comparison described in paper excerpt (specific sample size and statistical details not provided in the excerpt).
Consumers are increasingly delegating purchase decisions to AI agents, providing natural-language descriptions of their preferences and identity.
Asserted in paper's introduction/abstract as a background trend; no empirical sample or citation provided in the excerpt.
Expert validation established strong relevance and practical utility for the framework, with a mean score of 4.6/5.
Structured validation exercise with five domain experts in AI ethics, corporate governance, and fintech regulation; paper reports the mean validation score as 4.6/5.
Analysis revealed four foundational governance pillars: Accountability, Transparency, Fairness, and Compliance.
Theme extraction from the SLR of 45 peer-reviewed publications (2022-2025) reported in the paper; these four pillars are presented as the core components of the proposed framework.
The study develops and validates an integrated conceptual framework that incorporates corporate governance principles with mechanisms for algorithmic fairness to foster ethical outcomes in SME fintech lending.
Two-phase research approach described in paper: (1) systematic literature review (45 peer-reviewed publications, 2022-2025) and (2) structured validation with five domain experts in AI ethics, corporate governance, and fintech regulation.
AI-driven credit assessment platforms promise greater efficiency in fintech lending.
Statement in paper (conceptual claim); supported by related literature cited in the SLR of 45 papers but no empirical efficiency metric reported in this paper.
The rapid growth of fintech lending has reshaped financial access for SMEs through AI-driven credit assessment platforms.
Assertion in paper's background; positioned as established context for study (no specific empirical estimate given). The paper's SLR (45 peer-reviewed publications, 2022-2025) is presented as the literature basis for context.
To a lesser extent, fears of AI automation drive demand for schemes that guarantee income regardless of employment status.
Findings from the 2024 OECD 'Risks that Matter' survey reported in the paper (survey-based measure of support for income-guarantee schemes conditional on fear of automation).
Rather than increasing support for traditional interventions such as unemployment benefits and training programs, these fears primarily drive demand for measures that preserve the social role of work and protect it from automation, such as robot taxes.
Results from the 2024 OECD 'Risks that Matter' public opinion survey analyzed in the paper (survey-based association between fear and policy preferences).
Make is most compelling for commodity utilities and for differentiating custom applications in the AI era.
Paper's typology and normative recommendation derived from conceptual analysis (no empirical validation reported).
AI fundamentally transforms the governance properties of the Make option, shifting it from Williamson's pure hierarchy to a hybrid governance form that combines code ownership with external AI infrastructure dependency.
Conceptual argument combining transaction cost economics, resource-based view, and assessment of AI infrastructure characteristics (no empirical testing reported).
The 'SaaSocalypse' narrative predicts that AI will render large segments of the Software-as-a-Service market obsolete by enabling firms to build software in-house at a fraction of historical cost.
Statement summarizing an extant narrative in industry and literature (paper cites/describes this narrative; no empirical test in the paper).
Advances in generative artificial intelligence, particularly agentic coding systems capable of autonomous software development, are disrupting the economics of the make-or-buy decision for enterprise applications.
Paper's conceptual analysis combining transaction cost economics, resource-based view, and assessment of current AI capabilities (no empirical sample reported).
Grounding recommendations in validated research offers leaders a framework for navigating AI's labor implications responsibly.
Paper asserts that its synthesis and recommendations provide a practical framework for leaders; no empirical validation of the framework is reported in the abstract.
Evidence-based organizational responses (transparent workforce planning, skills investment, redesigned roles, adaptive governance, and long-term capability-building) can mitigate harm and prepare organizations for workplace transformation.
Paper proposes these organizational responses grounded in the synthesized empirical literature; this is a recommendation rather than an empirically tested intervention in the paper abstract.
There is an absence of a comprehensive national strategy in Israel for AI in employment, and the paper calls for the development of a forward-looking regulatory framework that balances innovation with protection of fundamental rights (dignity, equality, privacy), transparency, human oversight, and fairness.
Normative policy recommendation based on the paper's regulatory analysis; not an empirical finding and no policy-design experiments are reported in the excerpt.
The AI-driven transformation is accompanied by an increasing emphasis on reskilling and continuous learning, reflecting a shift from workforce replacement to reconfiguration of modes of employment.
Reported observation in the paper about workforce development trends; no quantitative measures of reskilling uptake or program counts are provided in the excerpt.
Israeli legal scholarship reflects broad interdisciplinary engagement with AI across labor law, intellectual property, privacy, constitutional law, and additional fields; the study advances theoretical models, including reconceptualizations of accountability, creativity, and the role of AI as a legal actor.
Literature review/academic survey and theoretical contributions reported in the paper; specific counts of publications or analytical methods not provided in the excerpt.
Israel is a leading “AI Nation,” characterized by exceptionally high levels of technological integration across both the private and public sectors.
Statement in paper based on the author's characterisation of national-level technological integration; specific empirical measures or sample size not provided in the excerpt.
Through targeted prompting inspired by these findings, we modify agents' negotiation behavior and improve win rates from 22.2% to 32.7%.
Intervention experiment reported in the paper where prompts were changed and resulting agent win rates were measured.
Governance maturity is therefore not merely a constraint on AI adoption; it is a condition that shapes whether capability improvements translate into productive deployment.
Synthesis/conclusion drawn from the analytical model showing governance affects the mapping from capability to productive deployment.
Governance investment that reduces breach-loss magnitude shrinks the paradox region itself.
Analytical model result showing how changes in governance (modeled as reductions in breach-loss magnitude) affect the parameter region where the deployment paradox occurs.
AI agents are now running real transactions, workflows, and sub-agent chains across organizational boundaries without continuous human supervision.
Statement in abstract describing observed industry trend; paper reports a structured survey of industry trends, emerging standards, and technical literature as its method for situating this observation.
Addressing AI in evaluative systems requires treating monitoring (AI detection) and loosened selectivity as complementary design instruments.
Policy implication derived from model results and constrained optimization of editorial policy in the post-transition regime; argued in the paper's conclusion.
The proposed approach reframes AI control from optimizing decisions to governing their admissibility, introducing a protocol-level abstraction that operates independently of model architecture or training methodology.
Conceptual argument and proposal in the paper asserting architecture-agnostic protocol abstraction. No empirical tests across architectures or training methods reported.
Through a scenario-based case study, we demonstrate how identical AI outputs can lead to divergent outcomes when evaluated under a Right-to-Act protocol, preserving reversibility and preventing premature or irreversible actions.
Scenario-based case study (illustrative demonstration). The paper reports example scenarios rather than empirical experiments; no sample size or quantitative evaluation reported.
Unlike compensatory systems, where high-confidence signals can override failed conditions, the proposed framework enforces strict structural constraints: if any required condition is unmet, execution is halted or deferred.
Conceptual distinction and protocol rule specification in the paper (formal description of non-compensatory enforcement). No empirical testing reported.
We introduce the Right-to-Act protocol, a deterministic, non-compensatory pre-execution decision layer that evaluates whether an AI-generated decision is permitted to be realized at all.
Proposed method / conceptual contribution and formal definition provided in the paper (formalization and protocol specification). No empirical validation or sample size reported.
That compliance layer can improve oversight by making departures from law easier to detect.
Claim supported by the paper's analytical argumentation (no empirical evidence reported).
For probabilistic AI to be incorporated into public administration it must be embedded in a compliance layer that makes decisions reviewable, repeatable, and legally defensible.
Stated as a normative/architectural claim in the paper; supported by conceptual argument rather than empirical testing.
Governments are increasingly interested in using AI to make administrative decisions cheaper, more scalable, and more consistent.
Stated as background motivation in the paper (no empirical data or sample size reported).
LLMs reveal their ability to approximate adaptive equilibria beyond static mechanism design.
Interpretive claim based on simulation results showing LLM bidders adapt and achieve favorable outcomes when mechanism assumptions fail (e.g., static budgets). This is an inferential claim drawn from comparative experiments described in the paper; no numerical quantification provided in the excerpt.
When theoretical assumptions break—such as under static budget constraints—LLMs sustain longer participation and achieve higher utilities.
Reported simulation results under scenarios violating VCG truthfulness assumptions (example: static budget constraints). The paper states LLM bidders maintained participation for longer and obtained higher utility than truthful and heuristic strategies in these scenarios. No numeric sample sizes or quantified effect sizes provided in the abstract.
Unlike heuristics, LLMs leverage historical outcomes and prompt-based reasoning to adapt their bidding behavior dynamically.
Method description and reported behavioral difference: the paper states LLM bidders incorporate prior auction outcomes and prompt engineering to inform bids, contrasted with static heuristic strategies. Based on simulation implementation rather than field deployment; no sample size provided.
Technological capability (AI) and board diversity are complementary in strengthening corporate governance and fiscal discipline in developing economies.
Synthesis/interpretation of empirical results (main effects and interaction effects) from panel regressions and robustness analyses on 1,586 firms across 2009–2023.
The main findings are robust to alternative tax avoidance measures, alternative BGD specifications, heterogeneity analyses, and selection-bias corrections (Heckman, propensity score matching, and instrumental-variable 2SLS approaches).
Reported robustness checks in the paper applying multiple alternative variable specifications and methods for selection-bias correction on the primary sample.
AI capability significantly strengthens the relationship between BGD and effective tax rates; firms with higher AI adoption exhibit a stronger governance effect of gender-diverse boards on tax compliance.
Interaction models estimated on the same balanced panel (1,586 firms, 2009–2023) using lagged AI capability specification; estimated with firm FE and dynamic two-step System GMM, with reported statistically significant interaction effects.
Board gender diversity (BGD) is positively associated with effective tax rates, implying lower levels of corporate tax avoidance.
Empirical analysis using a balanced panel of 1,586 non-financial firms from developing economies over 2009–2023; firm fixed effects models and dynamic two-step System GMM estimations used to address unobserved heterogeneity, endogeneity, and persistence of corporate tax behavior.