Evidence (1222 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
21267 claims
Filter claims →
Productivity
17978 claims
Filter claims →
Governance
17038 claims
Filter claims →
Human-AI Collaboration
16914 claims
Filter claims →
Org Design
11104 claims
Filter claims →
Innovation
11087 claims
Filter claims →
Labor Markets
6711 claims
Filter claims →
Skills & Training
5616 claims
Filter claims →
Inequality
4343 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 | 1880 | 496 | 296 | 1854 | 4721 |
| Organizational Efficiency | 2906 | 665 | 438 | 180 | 4210 |
| Governance & Regulation | 2162 | 929 | 480 | 247 | 3866 |
| Technology Adoption Rate | 1533 | 545 | 278 | 210 | 2593 |
| Decision Quality | 1391 | 534 | 321 | 173 | 2429 |
| Output Quality | 1298 | 472 | 231 | 145 | 2153 |
| AI Safety & Ethics | 682 | 821 | 230 | 90 | 1837 |
| Research Productivity | 855 | 253 | 121 | 425 | 1675 |
| Firm Productivity | 1105 | 171 | 175 | 73 | 1531 |
| Task Allocation | 735 | 229 | 361 | 99 | 1433 |
| Market Structure | 457 | 461 | 251 | 47 | 1222 |
| Innovation Output | 673 | 94 | 108 | 36 | 913 |
| Task Completion Time | 499 | 118 | 43 | 38 | 702 |
| Firm Revenue | 458 | 130 | 61 | 26 | 677 |
| Skill Acquisition | 381 | 122 | 113 | 34 | 650 |
| Consumer Welfare | 316 | 176 | 115 | 39 | 648 |
| Employment Level | 223 | 143 | 177 | 53 | 600 |
| Error Rate | 246 | 282 | 44 | 19 | 594 |
| Fiscal & Macroeconomic | 283 | 142 | 78 | 52 | 562 |
| Inequality Measures | 103 | 329 | 106 | 13 | 552 |
| Worker Satisfaction | 225 | 185 | 63 | 30 | 503 |
| Automation Exposure | 158 | 155 | 72 | 37 | 426 |
| Regulatory Compliance | 186 | 126 | 35 | 14 | 362 |
| Team Performance | 193 | 56 | 51 | 24 | 326 |
| Developer Productivity | 224 | 58 | 27 | 13 | 323 |
| Wages & Compensation | 148 | 108 | 50 | 17 | 323 |
| Training Effectiveness | 218 | 44 | 21 | 27 | 313 |
| Job Displacement | 23 | 159 | 53 | 5 | 240 |
| Hiring & Recruitment | 109 | 61 | 32 | 11 | 215 |
| Skill Obsolescence | 16 | 107 | 26 | 6 | 155 |
| Creative Output | 71 | 44 | 28 | 6 | 150 |
| Social Protection | 58 | 31 | 12 | 3 | 104 |
| Labor Share of Income | 29 | 43 | 25 | 2 | 99 |
| Worker Turnover | 45 | 29 | 6 | 4 | 84 |
| Industry | — | — | — | 1 | 1 |
Customization can create more persistent competitive differentiation through proprietary data, task-specific capabilities, and integrated workflows, but the resulting economic returns depend on firms' ability to manage technical, organizational, and regulatory frictions.
Cross-industry comparisons, qualitative interviews, and comparative cases; the paper is described as exploratory and does not establish a causal effect of customization on firm performance.
AI-enabled operational advantages, including improved dispatch and bidding strategies, could alter competition and increase market-power concerns if advanced capabilities become concentrated among a few firms.
Conceptual economic implication based on AI-enabled dispatch and bidding; no market concentration estimate or empirical competition analysis is reported.
Bear markets are the state in which AI-related assets most effectively influence sustainable investment opportunities, whereas in normal and bull markets AI-related assets are comparatively more likely to receive shocks.
Quantile-VAR analysis of directional spillovers across bear, normal, and bull market states.
LLM agents predominantly improved standing orders in small $0.01 increments, preserving potential profit but increasing the risk that trades would not occur before the round ended.
The authors analyze individual order improvements across the simulated experiments and report the empirical distribution of improvement sizes. The supplied text does not state the exact number of observations classified as $0.01 increments.
The paper argues that the observed language and location effects are better characterized as a market-selection gate than as a simple preference for particular brands, but it does not identify the mechanism implementing the gate.
Observed cases where the interface explicitly mentioned national VAT or e-invoicing constraints but still named only foreign products; authors acknowledge that retrieval, instruction tuning, and list construction remain consistent explanations.
Russian, a minority language on the Estonian connection, produced an intermediate localization pattern: Estonian suppliers appeared in 4 of 6 runs, compared with 0 of 6 in English and local suppliers appearing in every Estonian-language run.
Interleaved six-run comparison of Estonian, Russian, and English prompts from the same residential Tallinn connection.
AI can lower entry costs for masstige positioning, potentially increasing competition while also allowing incumbent platforms to entrench winners through personalisation and data advantages.
Conceptual analysis of platform economics, AI-enabled marketing, and data advantages; no market-entry or concentration estimates are reported.
Fragmented regulatory approaches, including differences between strict data-localization and permissive regimes, hamper global AI markets, whereas multilateral coordination and interoperable standards support scale economies.
Policy analysis of regulatory fragmentation and harmonisation, extended to data governance and digital-service regulation affecting international AI markets.
Supplier channel exit disrupts the inherent zero-sum competitive relationship between the supplier and the MCN, with equilibrium outcomes varying according to the combination of cross-channel traffic-spillover intensity and consumer preference for AI streamers.
Two-stage Stackelberg game model comparing long-term dual-channel cooperation with short-term cooperation followed by supplier exit; equilibrium analysis under different parameter combinations.
The platformisation and datafication of education are reshaping educational markets, data governance, and pedagogical practices while introducing new power asymmetries based on surveillance and data monetization.
Conceptual analysis of EdTech and platform actors, drawing on existing scholarship and contemporary developments in technology and education.
Algorithmic ranking and recommendation can allocate attention in ways that create distributional externalities for news producers and affect demand for different types of content.
Theoretical implication for platform economics and attention markets; the article does not quantify the externalities or estimate effects.
Incorporating constitutive dynamics into matching theory implies that worker preferences and suitability may be endogenous and shaped by organizational experience rather than fully stable and observable in advance.
Theoretical implication for labor-market matching and market-design models; no equilibrium model or empirical estimate is presented.
Rapid technological change in diagnostics, pharmaceuticals, and digital health is shifting which countries produce key health inputs.
Qualitative policy analysis based on observed technological diffusion and changes in production capacity.
Following the COVID-19 shock, European airlines converged in some network dimensions while remaining persistently differentiated in others.
The post-pandemic longitudinal comparison evaluates OT-based network dimensions across the 32-airline sample.
Relative to a 2016 baseline, the longitudinal analysis identifies diverse adjustment trajectories across airlines.
Airline-year distributions are compared with a 2016 baseline over the 2016–2025 observation window.
In an application to 32 European airlines observed from 2016 through 2025, airline similarities were multidimensional and varied by carrier type and network feature.
The empirical analysis uses scheduled passenger-flight data for 32 European airlines over the 2016–2025 period and compares carriers across multiple OT dimensions.
Widespread adoption of similar explainable ESG models could increase herd behavior or model correlation across financial institutions, creating ambiguous systemic-risk effects.
Conceptual risk assessment in the paper's AI-economics implications; the supplied text reports no empirical test of systemic effects.
In the simulations, temporal dynamics arise from repeated interactions among bidding strategies and feedback from market clearing rather than from exogenous changes in demand, marginal costs, or network parameters.
The experimental setup holds exogenous fundamentals fixed within each scenario while agents repeatedly bid and receive market-clearing feedback.
The simulations use a stylized seven-bus electricity market with four strategic generators, inelastic nodal demand, transmission constraints, and locational marginal pricing.
Description of the simulated market setup and strategic participants.
The experimental design contains 18 electricity-market environments formed by a full factorial combination of network representation, demand level, and marginal-cost heterogeneity.
The paper specifies two network conditions, three demand levels, and three marginal-cost heterogeneity levels, yielding 2 × 3 × 3 = 18 environments.
The authors argue that supra-competitive profits alone are insufficient to establish tacit collusion, because elevated profits can arise without collusive coordination.
Conceptual analysis motivating the use of multiple collusion criteria rather than profit comparisons alone.
In AI markets, entrepreneurial choices such as open-source versus proprietary development, licensing, and platform access controls will shape the excludability of AI outputs and the market structure of AI industries.
Application of the paper's conceptual framework to AI commercialization; presented as an implication or prediction rather than an empirical finding.
Property-rights entrepreneurship can produce new markets, shifts in firm boundaries, reallocation of rents, and institutional change.
Conceptual analysis of the potential outcomes associated with product, asset, and resource rights.
Disclosure requirements may increase firms' investment in documentation and fairness-aware development while also creating entry barriers for smaller firms unless compliance support is provided.
Conceptual analysis of firm incentives and market-structure implications; no measured compliance costs or entry effects are reported.
The study introduces reputational adjacency, describing reputational pressure on a firm caused by a competitor’s technological success rather than by failure of the focal firm.
Conceptual development based on the DeepSeek R1 case and observed stakeholder judgment shifts.
Improved recommendation technology shifts played-title consumption away from superstar titles toward middle-tail and long-tail titles.
The paper states that the experiment tests treatment effects on play shares across title buckets, following the observed redistribution of recommendations.
Improved recommendation technology shifts recommendation share away from superstar titles primarily toward middle-tail titles, with quantitatively small changes for long-tail titles.
Treatment effects on within-user recommendation shares by title bucket, estimated using a randomized treatment indicator; results are shown in Figure 2a and supported by more granular ventile analyses.
The study finds that specialist and generalist sponsors differ across sector selection, company-level entry pricing, acquisition consideration, and short-term valuation outcomes.
Analysis of 808 technology-unicorn acquisitions announced between 2010 and 2024, with 601 U.S.-based acquisitions in the primary analysis and 207 international acquisitions in an extension.
Blockchain may change audit-market structure by enabling technical entrants such as cybersecurity and smart-contract-auditing firms to compete with traditional auditors, while encouraging incumbents to offer system-assurance services.
Conceptual analysis of how blockchain-related assurance capabilities could affect competition, service bundling, and auditor business models.
Entity List sanctions reconfigure rather than sever global academic collaboration networks.
The study combines coauthorship-network analysis with evidence of increased ties to non-listed Chinese intermediaries and no significant decline in U.S. collaboration.
Faster diffusion of AI in production technologies could alter comparative advantage, reshape supply chains, and increase divergence in technical standards and regulatory regimes.
Strategic and geopolitical implications inferred from the paper's diffusion-forward framework; no quantified cross-country estimate is reported.
The central axis of AI competition should be reframed from achieving AGI or frontier-model milestones first to integrating AI rapidly, broadly, and effectively into productive activity.
The paper's comparative strategic argument, supported by policy analysis and observed registration and patent activity.
When competing estimates are noisy, value-of-information reasoning can increase the strategic specialist’s utility at the expense of other specialists.
The decentralized experiments compare strategic utility and outcomes for other specialists under noisy competing estimates.
The paper argues that local inference broadens participation at the point of execution while relocating value capture and control into the infrastructure required to make models runnable.
Mixed-methods synthesis of pull-request patterns, repository discussions, corporate statements, contributor blogs, hardware-backend integrations, model-conversion work, and Hugging Face's absorption of the project.
Omitted-effort semantics differ across model IDs, including among models from the same provider.
A dated census of seven current-menu model IDs, official documentation, Models-API metadata, and 19 preregistered single-item contract probes examined accepted requests, served identity, echoed effort, thinking structure, token details, and disabled-thinking behavior.
Coordination of standards and data governance can influence market formation and competitive dynamics, potentially reducing concentration by lowering entry barriers or limiting incumbent control over key data and standards.
Theoretical economic implication concerning regulatory coordination, market structure, and competition; no empirical market analysis is reported.
Generative artificial intelligence changes the basis of comparative advantage in international services trade rather than eliminating comparative advantage.
The paper's theoretical framework combining comparative advantage theory, trade-in-tasks theory, and automation research.
Large buyers with stronger text-analytics capabilities may gain bargaining power and supplier lock-in advantages, potentially increasing buyer-side market power.
Conceptual market-structure implication based on superior monitoring, matching, and cumulative data advantages; no market-level data or causal estimate is reported.
Platform ranking and recommendation algorithms materially shape the returns to DFC and thereby affect market equilibria and creator bargaining power.
Theoretical framework connecting platform design choices, algorithmic amplification, DFC returns, and market structure; empirical support is illustrative rather than causal.
Large upfront and continuing investments in data collection, storage, compute, and long-term maintenance may favor well-resourced actors and contribute to concentration around foundation-model providers.
Economic reasoning about fixed and recurring infrastructure costs and economies of scale in Earth-AI development.
Using multimodal data in financial models may change the speed and accuracy with which markets incorporate information and may generate return predictability relevant to tests of semi-strong market efficiency.
Research agenda and implications discussion concerning multimodal cues, asset pricing, event studies, and market information incorporation.
When AI adoption is driven by demand factors such as lower fixed adoption costs, the relationship between AI diffusion and the average markup is non-monotonic.
Model comparative statics under changes in the fixed cost of AI adoption.
The non-monotonic relationship between AI diffusion and industry concentration persists even when firms can combine AI with their own data.
Model extension allowing firms to complement AI with proprietary data, followed by comparative-static analysis.
The model's turning point for the relationship between AI diffusion and industry concentration occurs at approximately a 15 percent AI adoption rate.
Comparative-static results from the calibrated general-equilibrium model.
AI diffusion has a hump-shaped (non-monotonic) relationship with industry concentration: concentration initially rises as AI adoption increases, then falls once AI adoption becomes sufficiently widespread.
General-equilibrium model with heterogeneous firms, fixed AI adoption costs, variable markups, endogenous AI production, and calibrated comparative-static exercises.
Industrial policies supporting domestic AI and defense supply chains are likely to alter competition, raise barriers to entry, and concentrate capabilities in national champions.
Forward-looking economic inference concerning subsidies, strategic firms, domestic supply chains, and rapid-prototyping capacity; no firm-level or market-level data were presented.
Firms sort into three distinct clusters based on their institutional-pressure profiles.
Cluster analysis of panel data from Chinese A-share listed firms covering 2007–2022 identified three groups with distinct combinations of institutional pressures.
Cross-system agreement on the top 20 recommended venues was low, with Jaccard similarity ranging from 0.33 to 0.54.
Comparison of the top-20 recommendation sets produced by ChatGPT, Claude, Gemini, and Perplexity.
Enterprise adoption of AI-driven revenue systems may alter market power, consumer surplus, and competitive outcomes through mechanisms such as dynamic pricing and targeted offers.
Policy and competition implications proposed by the article; no causal estimates or empirical effect sizes are reported.
After independently trained policies are rematched, maximum betweenness and attractor in-degree retain their predicted signs and correlations with the Collusion Index, while basin fraction and the number of attractors become informative when competition fragments the state space.
Robustness test rematching independently trained policies that were overfit to their original training rivals.