Evidence (1809 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 |
The AAITCF treats context as constitutive of intervention effectiveness and highlights underexplored causal pathways from AI deployment to long-term institutional change, taxpayer trust, and equitable fiscal governance.
Framework description and identification of research gaps in the paper based on the literature synthesis using the CIMO framework.
The study finds that prior reviews tended to focus narrowly (e.g., on detection metrics, behavioral dynamics, or ethical deficits) without integrating institutional boundary conditions, governance capacity, or an overarching theoretical framework.
Critical comparison and gap analysis of existing review literature as reported in the paper's introduction and synthesis sections.
Policy implications derived from the literature include interventions spanning labor transition (reskilling/transition support), competition regulation, and digital governance.
Narrative synthesis of policy recommendations across the 78 studies and institutional reports included in the SLR.
AI policies' carbon outcomes depend on regional economic structures, implying the need for spatially differentiated governance.
Interpretation/implication drawn from heterogeneous and spatial analyses showing region-specific effects; result is policy recommendation based on study findings (supporting analyses referenced but not detailed in abstract).
Heterogeneous effects: emissions decreased in the Pearl River Delta and increased in the Chengdu–Chongqing region and in resource-based cities (these heterogeneous findings are statistically marginal).
Subgroup/regional heterogeneity analysis comparing policy effects across regions (Pearl River Delta, Chengdu–Chongqing, resource-based cities); statistical significance described as marginal in the paper (no sample sizes or exact p-values provided in abstract).
The research is limited by the current state of AI technology and the available proxies; therefore the validity of the present optimistic findings must be continually re-evaluated.
Authors' stated limitations in the abstract noting rapid AI advancement and proxy measurement constraints.
We propose 'contextuality' — the degree to which an AI system autonomously accesses a user's accumulated knowledge capital — as a dimension of AI-mediated inequality that complements, but is not reducible to, the Sharp et al. framework.
Conceptual proposal and definitional contribution in the paper presenting contextuality as a new analytic dimension.
Sharp et al. (2025) introduce "agentic inequality" as a framework for analyzing disparities in access to AI agents across three dimensions: availability, quality, and quantity.
Statement and citation in the paper (reference to Sharp et al. 2025); descriptive synthesis of prior work.
The paper identifies four systemic tensions generated by embodied AI adoption: openness versus control; scaling versus local fit; automation ambition versus reliability constraints; and monetization versus trust.
Explicit listing of four tensions in the abstract as theoretical findings (conceptual analysis).
Data generated through physical use of embodied AI travels beyond the adopting firm (i.e., data flows cross firm boundaries).
Explicit conceptual claim in the abstract about data movement across ecosystems (theoretical observation).
Embodied AI implies a double learning loop: a closed learning loop inside the adopting firm (transforming situated use into operational feedback and workflow changes) and an external learning loop across the ecosystem of technology providers, component suppliers, software firms, platform orchestrators, and users.
Conceptual model/argument presented in the abstract describing intra-firm and inter-organizational learning loops (theoretical development).
Because AI externalities differ in nuanced ways, tax policy must be carefully designed and matched to the specific harms and policy objectives.
Author conclusion/recommendation based on the paper's analysis of heterogeneous AI externalities and tax instrument trade-offs; normative claim in text (no empirical test in excerpt).
The benefits and pitfalls of these instruments include feasibility, measurement problems, incidence, leakage, and innovation costs.
Author assessment summarized in paper identifying common advantages and disadvantages of proposed tax instruments; descriptive/theoretical evaluation rather than empirical evidence in the excerpt.
Possible tax instruments for AI include corporate income and rent-based taxes, consumption taxes on AI-related services, and excise instruments tied to specific AI activities.
Author survey of tax instruments presented in the paper; descriptive listing rather than empirical claim (paper states these instruments are discussed/surveyed).
The mandate acted as a catalyst rather than a direct driver: because adoption and usage intensity were not randomly assigned, the evidence strongly implicates an adoption-and-use channel rather than exact causal attribution.
Authors' methodological caveat based on observational (non-randomized) adoption and usage intensity; interpretation of DiD estimates as indicative of channels rather than definitive causal estimates.
Optimal tax and regulatory policies that achieve Pareto-improvements differ depending on whether there is competition in AI production.
Policy analysis within the theoretical model deriving optimal tax/regulatory prescriptions under different market structures (competitive vs monopolistic). No empirical sample reported.
Rather than proposing a deterministic growth model, the study advances a conditional and ecosystem-centered interpretation of AI-led development.
Authors' interpretive conclusion based on their structured review and the integrative innovation-ecosystem framework synthesizing mechanisms and contextual dependencies in the 2015–2025 literature.
The proposed model demonstrates how natural resource dynamics, financial systems, and AI technologies form an interdependent triadic structure in which disturbances in one domain propagate across the entire system.
Presentation of a conceptual/formal model (systems analysis) in the paper showing interdependencies; no empirical dataset or sample size provided.
The research conceptualizes sustainability as a nonlinear adaptive process characterized by dynamic feedback loops and emergent systemic behavior.
Theoretical/systems analysis and conceptual argumentation in the paper; no empirical validation or sample size reported.
Better measurement matters, but improved measurement alone will not close the coordination gap between researchers and policymakers.
Authors' analytical conclusion arguing that measurement improvements are necessary but insufficient.
The paper develops the concept of 'bidirectional dynamics' in digital sovereignties, applying a paradoxical view to interpret institutional control objectives and individual autonomy aspirations as persistent organizational tensions in AI adoption.
Theoretical/conceptual development grounded in the empirical single-case study; concept introduced and motivated by observed tensions in the organization (empirical method details and sample size not provided).
Early digital transformation presents tensions but also synergies between digital sovereignty levels in AI adoption.
Empirical observations from the single-case study of a Nordic public transportation organization during early AI adoption; qualitative examples and analysis (specific methods/sample size not stated).
Generative engine optimization (GEO) should be studied not only as a security risk, but also as an emerging marketing practice that shapes market competition.
Paper's concluding/interpretive statement based on the experimental findings about LLM recommendation dynamics and GEO effects on brand recommendations.
A third possibility — the collective and self-organized stewardship of AI-relevant resources by communities (commons-governed approaches) — remains comparatively under-theorized in scholarship even as it proliferates in practice (e.g., data trusts, cooperatives, federated learning consortia, public compute initiatives, open-weight collaborations, community data sovereignty regimes).
Comparative literature review noting fewer theoretical treatments of commons approaches alongside cited examples of practical manifestations (lists of existing initiatives and models).
Human values produce societies that thrive or fail on the merits of those values — from failed states and extreme inequality to declining happiness, political polarization, and government dysfunction in the world's wealthiest democracies.
Descriptive/causal claim asserted by authors linking values to a range of societal outcomes; no specific empirical studies or samples cited in the abstract.
Forms of resistance exist, including localisation efforts and Indigenous ethical frameworks, but they remain structurally limited.
Synthesis of examples and themes across the 50 reviewed articles noting reported resistance strategies and their limits.
While net-zero targets for 2050 may be achieved, critical emission risks may appear in intermediate years and the EU may compromise its carbon‑neutral goals unless policies adapt to the accelerating digital transformation.
Scenario trajectories from the optimisation model indicating that 2050 net-zero remains attainable in some scenarios but with interim emissions overshoots; policy conclusions drawn by the authors.
The economic consequences of generative AI in financial markets depend critically on institutional context (regulatory and governance capacity).
Synthesis of heterogeneous treatment effects and interaction results across markets with varying governance/regulatory quality in the cross-market panel analysis.
The paper characterises the Glassbox architecture and grounds it in a benefit eligibility scenario, identifying foundational challenges — semantic alignment, dynamic model construction, probabilistic grounding, and human governance — that must be solved to realise it at scale.
Descriptive summary of the paper's contributions and identified research/engineering challenges; based on the authors' conceptual analysis and scenario exposition.
These examples show an important shift in the governance of wealth chains – the creation of new forms of infrastructural power through which algorithmic models may become central nodes in tax governance.
Synthesis/interpretive conclusion in the abstract that the illustrative examples imply a governance shift and new infrastructural power; presented as interpretive argument rather than empirically demonstrated in the abstract.
This signals a transformation of the assumed information asymmetries between suppliers, clients, and regulators that sits at the heart of the Global Wealth Chains framework.
Conceptual claim in the abstract linking technological change to shifts in information asymmetries within the Global Wealth Chains framework; presented as interpretive argument rather than supported by reported empirical data in the abstract.
A key development is a move away from deliberate opacity for secrecy purposes into systems that search for the optimal exploitation of legal affordances.
Analytic/interpretive claim made in the abstract about a shift in practices; presented as an argument based on the authors' reflection and examples rather than empirical measurement in the abstract.
Grounding the concept of defensive AI governance in organisation-level evidence from the Global South contributes to debates on platform power, journalistic agency, and AI governance in journalism.
Theoretical/interpretive claim based on the study's case of Al-Masry Al-Youm and its empirical insights; presented as a contribution to scholarly debates. Sample size not reported in the excerpt.
The authors introduce the concept of 'defensive AI governance' to describe how AI adoption is managed through organisational practices of limitation, supervision, and infrastructural self-protection.
Conceptual contribution grounded in organisation-level qualitative evidence from interviews and analysis of Al-Masry Al-Youm's practices; the concept is derived from the study's empirical findings. Sample size not reported in the excerpt.
Human and algorithmic actors jointly influence strategic outcomes, motivating the concept of 'hybrid upper echelons' in which executive influence increasingly shifts from making decisions to configuring and governing AI-enabled decision processes.
Theoretical contribution based on integration of management and IS literature in the concept-centric review; proposition of a new conceptual framework ('hybrid upper echelons') rather than primary empirical validation.
These behavioral differences have implications for deployment of agentic AI in scientific computing workflows, such as trade-offs between speed versus auditability, silent versus transparent error handling, instruction interpretation, and the criticality of intermediate data representations in multi-model pipelines.
Authors' discussion and interpretation based on observed experimental differences between the two agents across the runs.
The limitations in the audit reports reflect symbolic compliance (per institutional theory), while stewardship theory highlights potential for deeper accountability.
Theoretical interpretation using institutional theory and stewardship theory presented in the paper (argumentative rather than empirical).
Adaptive governance conditions how AI-driven capabilities translate into sustainability and risk outcomes.
Comparative analysis across the three jurisdictions (China, US, UK, 2022–2025) integrating quantitative indicators and qualitative documentary evidence, with the abstract highlighting the 'conditioning role of adaptive governance'.
Comparative analysis of Japanese, European, and United States legal frameworks shows differing treatments of translation data and points toward the need for redistributive design to remedy unequal attribution and capture.
Comparative legal analysis across jurisdictions (Japan, EU, US) and normative argument proposing redistributive design directions; no experimental or quantitative evaluation provided.
AI-induced workforce disruption is not only a labor market issue but also an enterprise governance challenge.
Argument/position advanced in the paper highlighting governance responsibilities for firms implementing AI.
Drawing on the partial equilibrium model of Gries and Naudé (2022), existing economic frameworks may inadvertently overlook these factors.
The paper's theoretical critique referencing Gries & Naudé (2022); argument is based on model comparison and conceptual analysis rather than new empirical tests.
These findings have broader implications for productivity, equity, and capacity across the global research system.
Discussion/interpretation in paper based on causal results from randomized experiment; inference from observed behavioral changes and heterogeneous effects.
Recent Chinese regulatory initiatives addressing anthropomorphic and emotionally interactive AI services illustrate emerging governmental responses to the social and psychological risks associated with relational AI.
Cited as an illustrative example in the recommendations; the text references Chinese initiatives but does not provide specific citations, legal texts, or empirical evaluation within the document.
Regulatory approaches to advanced AI systems are evolving differently across major jurisdictions.
General observation in the recommendations; no cross-jurisdictional comparative analysis or dataset provided in the text.
Algorithmic authority may both strengthen and undermine legitimacy of decisions in AI-enabled organizations.
Theoretical analysis in the paper presenting dual possibilities for algorithmic authority's impact on legitimacy, supported by conceptual reasoning and literature (no empirical test reported).
Verification cost and responsibility transferability determine whether the execution and accountability boundaries can move together.
Propositional/theoretical argument within the capability-level theory; supported by conceptual reasoning and illustrative cases, not by empirical estimation.
Comparative analysis reveals significant institutional differences between EU and Ukrainian legal systems that are relevant to regulatory stability, the cost of innovation, data accessibility, the balance of market power, and guarantees for consumers and employees.
Qualitative comparative examination of institutional and cultural/procedural differences between EU and Ukraine as presented in the paper (method: comparative approach; no quantitative metrics provided).
Most Ukrainian laws relevant to the digital economy are based on existing legal structures and systems, and Ukraine currently lacks a unified regulatory system specifically designed for artificial intelligence.
Comparative analysis of Ukrainian and EU legal frameworks as described in the paper (method: comparative approach; legal document review referenced qualitatively).
The study evaluates contemporary mitigation frameworks for algorithmic bias in HR settings.
Statement of the paper's evaluative aim; implies review/assessment of mitigation strategies but no specific methods or metrics provided in excerpt.
Techno-sovereignty is a mode of authority grounded in control over data, computation, and AI infrastructures, exercised through state, corporate, and community or Indigenous configurations.
Conceptualization and normative-theoretical analysis drawing on political theory and community/Indigenous approaches (qualitative, no quantitative data).