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 |
While several jurisdictions (UK, US, EU, India) have attempted to regulate gig work, most regulatory responses remain incomplete and fail to fully address platform accountability.
Comparative policy/regulatory analysis of the United Kingdom, United States, European Union and India assessing statutes, litigation and policy measures; qualitative assessment rather than statistical evaluation (no quantitative sample size reported).
Platform companies rely on contractual misclassification, corporate structuring, and the legal fiction of neutrality to separate control from liability.
Legal and corporate-structure analysis across jurisdictions, examining contracts, corporate forms and legal doctrines; based on comparative statutory and case-law review (no quantitative sample size reported).
The platform economy produces a deeply unequal labour structure marked by algorithmic control, economic dependency, surveillance, and lack of social protection.
Synthesis and critical analysis combining literature, policy review and comparative jurisdictional study to argue systemic effects on labour structure; primarily qualitative evidence and theoretical framing (no quantitative sample size reported).
Gig workers, though formally classified as independent contractors, are functionally subjected to pricing control, performance monitoring, automated penalties, and deactivation mechanisms that closely resemble managerial authority.
Descriptive/qualitative evidence in the paper: examples and analysis of platform design and management practices (algorithmic pricing, monitoring, penalties, deactivation); based on platform policy documents, case examples and comparative review (no quantitative sample size reported).
Digital labour platforms exercise employer-like control while avoiding employer-like legal responsibilities.
Argument and comparative legal analysis across jurisdictions (United Kingdom, United States, European Union, India) demonstrating platform practices and legal/regulatory responses; based on documentary/legal review and critical analysis (no quantitative sample size reported).
Shifts persist in even the newest AI models despite remarkable progress in AI modeling, post-training alignment and safeguards.
Asserted in paper; supported by later empirical validation across multiple models and production chatbots (see other claims), but no explicit sample size in this sentence.
ChatGPT-like AI behavior can shift, unnoticed, from desirable to undesirable (e.g., encouraging self-harm, extremist acts, financial losses, or costly medical and military mistakes), and no one can yet predict when.
Statement in paper framing the problem; qualitative observations and motivating examples (no numeric sample size provided in the excerpt).
These characteristics are properties of the tasks themselves rather than limitations of current AI models.
Conceptual argument in the paper asserting task-inherent properties drive resistance to automation; supported by theory and argumentation, not by empirical model-comparison experiments.
The resistance of Metis tasks to automation is not due to computational intractability but to institutional, social, and normative entanglements.
Theoretical argument differentiating computational from institutional/social/normative causes; supported by citations and cross-disciplinary theory rather than empirical causal identification.
There exists a class of entirely digital tasks, called 'Metis AI', that resist reliable AI automation.
Conceptual identification and definition introduced by the authors; supported by theoretical grounding in social sciences, philosophy, and humanitarian practice rather than empirical trials or quantified samples.
That digital-vs-physical framing misses the most consequential boundary: the one within digital tasks.
Normative/theoretical argument presented in the paper contrasting existing framing with a proposed alternative; grounded in cross-disciplinary literature rather than empirical measurement.
Severe penalties in underfunded Eastern systems, mediated by financial distress, drive families toward resource exhaustion.
Cross-country comparisons in SHARE-derived analyses showing larger financial penalties in underfunded Eastern European systems, with mediation analysis implicating financial distress and resultant resource exhaustion.
Financial distress acts as a profound multiplier of the burdens associated with palliative care.
Interaction/moderation analyses in SHARE-derived synthetic data showing that pre-existing financial distress amplifies financial and caregiving burdens under PC.
Socio-demographics heavily modulate exposure: lacking a spousal net inflates the burden.
Subgroup/moderation analyses in SHARE-derived data comparing households with and without spousal support, showing higher burdens when no spouse is present.
Non-cancer trajectories drive massive structural penalties that escalate at the distribution's tail, mechanically compounded by physical dependency.
Stratified analyses by disease trajectory (non-cancer vs cancer) using SHARE data (2016-2021) and quantile models showing larger penalties for non-cancer cases, especially in tail quantiles; physical dependency identified as a compounding factor.
Quantile treatment models expose a 'broken shield' for vulnerable households and severe tail events (PC protection fails or reverses at distributional tails).
Application of quantile treatment effect models to synthesized SHARE-derived digital twins (2016-2021), explicitly examining distributional/tail effects.
Parsing through LLM-generated code can be tedious and time-consuming, potentially negating the productivity gains promised by AI-coding tools.
Motivation/background statement in the paper: a qualitative claim about the cost (time/effort) of reviewing LLM-generated code; presented as motivation rather than empirically quantified evidence in the excerpt.
Employees experience technostress, anxiety and micro-political negotiation around AI tools in everyday work.
Reported experiences from semistructured interviews with 28 managers/professionals across 12 organizations; thematic analysis highlighting technostress and anxiety as themes.
An analysis of a 21-instrument inventory identifies an incentive gradient where geopolitical and industrial pressures systematically reward surface-level behavioral proxies over deep structural verification.
Empirical/qualitative analysis of an inventory of 21 governance instruments compiled and analysed in the paper (n=21 instruments).
Behavioural assurance, even when carefully designed, is being asked to carry safety claims it cannot verify.
The paper's normative and conceptual argument synthesising governance requirements and the epistemic limits of behavioural testing.
Current assurance methodologies (primarily behavioural evaluations and red-teaming) are epistemically limited to observable model outputs and cannot verify latent representations or long-horizon agentic behaviours.
Conceptual/analytic argument and review of existing assurance methodologies presented in the paper.
Overthinking is a shared and exploitable vulnerability in modern reasoning systems, underscoring the need for more robust defenses.
Conclusion drawn by authors based on their empirical findings described in the abstract (amplification of output length across multiple models and transferability experiments).
This overthinking behavior significantly increases inference latency and energy consumption, forming a potential vector for denial-of-service (DoS)-style resource exhaustion.
Authors assert increased latency and energy consumption as consequences of longer reasoning traces; framed as a potential attack vector in the abstract (no quantitative latency/energy measurements provided in abstract).
Large reasoning models (LRMs) exhibit a tendency to "overthink", producing excessively long and redundant reasoning traces when confronted with incomplete or logically inconsistent inputs.
Empirical observation reported by the authors based on experiments described in the paper (abstract references experiments across multiple SOTA reasoning models); no numerical sample size for inputs reported in abstract.
Distinct readability issue patterns and limited effectiveness of prompt engineering reveal a latent technical debt in LLM-generated code that could affect long-term maintainability.
Interpretation/conclusion in paper combining empirical findings (distinct issue patterns and limited prompt impact) to argue for potential technical debt and maintainability risks; presented as a forward-looking implication rather than a quantified causal estimate.
LLM-generated code displays distinct readability issue patterns compared to human-written code.
Empirical analysis of readability subcomponents/features showing different patterns of readability issues between LLM-generated and human-written code (paper reports qualitative/quantitative distinctions in issue patterns).
Policy responses in Europe are fragmented across the EU and Member State levels and do not match the potential scale of disruption from AGI.
Paper's policy analysis of EU- and Member-State-level responses (stated in abstract); no quantitative metrics provided in the abstract.
Europe has low rates of industrial AI adoption.
Paper's empirical/policy review claiming low industrial AI adoption in Europe (as stated in abstract); the abstract does not provide numeric adoption rates or sample sizes.
Europe exhibits structural weaknesses in compute infrastructure and talent retention.
Paper's structural assessment of Europe's AI value-chain capabilities (stated in abstract); no numerical measures provided in the abstract.
Europe has limited strategic awareness of frontier AI progress.
Paper's assessment of Europe's positioning based on policy analysis and review of capabilities monitoring (as stated in abstract); no supporting metrics or sample sizes provided in the abstract.
AGI could strain existing governance frameworks.
Paper's policy analysis describing potential mismatches between governance capacity and AGI-induced disruptions (as stated in abstract); no empirical tests or quantification reported in the abstract.
AGI could intensify interstate competition.
Paper's geopolitical analysis and scenario-based reasoning informed by trends in AI capabilities (stated in abstract); no quantitative measures reported in the abstract.
AGI could fundamentally alter the global distribution of economic and military power.
Paper's geopolitical analysis drawing on capability trends and scenario reasoning (as stated in abstract); no empirical quantification provided in the abstract.
Increased levels of AI assistance may degrade productivity, leading to potentially significant shortfalls under the model's identified conditions.
Model-based comparative-statics and steady-state analysis showing scenarios where marginal increases in AI assistance reduce expected task output; examples/parameter illustrations provided in the paper (theoretical, no empirical sample).
Introducing AI unreliability (errors/noise in AI outputs) in the model can also generate a productivity paradox: greater AI assistance may lower productivity.
Analytical/theoretical model incorporating AI unreliability; model derivations and examples demonstrating conditions under which unreliability leads to reduced productivity (no empirical data).
Incorporating endogeneity in skill development into the model can induce a productivity paradox where increased AI assistance reduces productivity.
Analytical/theoretical model of human-AI interaction with utility-maximizing human agents and endogenous skill development; steady-state and comparative-static analysis reported in the paper (no empirical sample).
Simulated users produce feedback dynamics that diverge from humans.
Temporal/interaction analysis in the replication showing differences in how simulators provide feedback across multi-turn interactions compared to humans.
Simulated users exhibit amplified position biases relative to human participants.
Behavioral comparison in the simulator replication showing stronger position biases in simulated responses than in human responses.
Simulated users discuss different topics compared to the human participants.
Analysis of conversation content in the simulator replication showing differences in topical distribution between simulators and humans.
Simulators perform far below human self-consistency baselines for individual judgements.
Comparison in the replication study between simulator consistency and human self-consistency on individual-level judgments; reported large performance gap (simulators far below humans).
Amplified sycophancy and relationship-seeking behaviours may introduce deleterious long-term consequences.
Authors' interpretation and cautionary note based on observed behavioral amplification after fine-tuning; presented as potential long-term risk rather than an empirically measured long-term outcome.
Existing AI-generated image detection benchmarks mainly evaluate standalone authenticity classification, cross-generator transfer, or forensic localization, leaving claim-conditioned fraudulent evidence detection underexplored.
Literature/contextual positioning in the paper contrasting prior benchmarks' focus with the proposed task.
There is a clear gap between generic AI image detection and reliable claim-conditioned refund-evidence verification.
Synthesis of experimental findings indicating that existing detectors and MLLMs are insufficiently reliable for the specific task of claim-conditioned refund-evidence verification.
Current MLLMs often recognize real-damaged evidence but fail on many fake-damaged subsets, with fake-damage detection rates (TPR) far below the 50% baseline on most generator subsets.
Experimental results reported in the paper comparing MLLM true positive rates (TPR) on real-damaged vs. fake-damaged subsets produced by multiple generators.
In a controlled experiment across six industry configurations (72 tool invocations using Qwen3-32B), unconstrained tool parameters produced a 43% hallucination rate for domain identifiers.
Controlled experiment reported in the paper: six industry configurations, 72 tool invocations, model used: Qwen3-32B; reported unconstrained parameter condition resulted in 43% hallucination rate for domain identifiers.
In multi-agent configurations the semantic training gap produces a compounding failure mode termed 'semantic drift'.
Analytical description and demonstration in the paper describing multi-agent interactions and observed/argued compounding failures (conceptual demonstration; no numeric sample stated).
The semantic training gap causes operationally incorrect outputs even when model responses are linguistically precise.
Demonstrations and examples reported in the paper showing cases where model outputs are linguistically fluent but operationally incorrect; supported by the paper's analysis and experimental illustrations (no numeric sample provided for this general claim).
There exists a 'semantic training gap': a structural disconnect between how AI systems acquire domain vocabulary through training and how manufacturing operations define meaning through ontological relationships.
Paper provides a formalization and conceptual framing of the gap (theoretical description and argumentation within the manuscript).
LLM-based AI agents deployed in manufacturing demonstrate statistical fluency with domain terminology but lack grounded understanding of operational semantics.
Stated assertion in the paper describing observed behavior of deployed LLM agents; supported by conceptual analysis and examples/demonstrations reported in the paper (no numeric sample size given).
Direct demographic targeting excludes users whose demographics the platform cannot infer ('unknown users') if advertising platforms do not provide a way to target unknown users directly, as is the case on Google Ads.
Platform capability statement about Google Ads (authors' description of Google Ads targeting options); no sample size provided.