Evidence (7560 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
Filtered →
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 |
Human Ai Collab
Remove filter
Acceleration in the Generate/Take Action phase translates into durable performance only when Analyze/Prioritize is de-biased by individuals and teams, and Measure/Review converts results into reusable knowledge with appropriate inference discipline.
Thematic conclusions from the 17 interviews and cross-case analysis (Gioia methodology) identifying conditional relationships across stages of the seven-stage growth pipeline.
GenAI enables small teams to expand capacity while creating new dependencies and coordination logics.
Empirical finding from 17 interviews indicating both expanded capacity and emergent dependencies/coordination needs.
GenAI drives structural recomposition across four domains: shifting roles, AI-embedded workflows, evolving capability expectations, and leaner work architectures.
Empirical finding from thematic analysis of 17 expert interviews reported in the results.
High-information AI improves short-run (immediate) performance without reducing post-AI outcomes on average in the experiments, but effects are heterogeneous across participants.
Experimental condition with high-information AI in the controlled logical reasoning task showing improved short-run performance and no average reduction in post-AI outcomes; heterogeneity in effects reported (sample size not provided in abstract).
Interpretability, trust calibration, and interface design matter, but they cover only part of what determines whether human-AI combination works.
Authors' argumentative claim based on their analysis and mapping of broader factors; presented as an evaluative conclusion rather than an empirical estimate.
Meta-analytic evidence shows moderate but heterogeneous effects of agentic/code-generation tools on productivity.
Reference to meta-analytic synthesis across studies reported in the paper (meta-analytic details not provided in abstract).
The benchmark therefore assigns value to coordination only when the corresponding performance, provenance, or representation claim is supported by explicit comparators.
Concluding statement in the paper tying value of coordinated AI agents to evidence from explicit baseline/comparator evaluations across performance, provenance, and representation dimensions.
For molecular sonification, the gain is representational rather than predictive.
Reported outcome for molecular structure to music task indicating improvements in representation/sonification quality but not in predictive performance.
The economics literature uses specific quantitative arguments and methods to estimate the changes produced by automation, and there is an ongoing debate in the field about these quantification methods.
Paper presents and synthesizes economic studies and methodological approaches (task-based methods, decomposition analyses, etc.) as part of a literature review and critical discussion.
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.
The paper analyses three primary vectors of AI bias in hiring: data bias, interaction bias, and evaluation bias.
Stated analytic framework in the paper (categorization of bias vectors); descriptive content rather than quantified empirical result.
This study examines the dual role of AI in the workplace: as a tool for bias reduction and as a potential vehicle for systemic discrimination.
Statement of the paper's research aim / framing; descriptive claim about the paper's scope rather than empirical finding.
AI alters strategizing practices (Strategy-as-Practice) by making strategy processes continuous and AI-augmented rather than episodic and purely human-driven.
Conceptual synthesis of Strategy-as-Practice literature; theoretical claim about process change to continuous, AI-augmented strategizing; no empirical sample.
AI redistributes resource control to stakeholders, challenging the Stakeholder Resource-Based View by changing who holds and controls strategically valuable resources.
Theoretical argument within the Stakeholder Resource-Based View stream; conceptual synthesis asserting redistribution of resource control to external stakeholders and algorithmic actors; no empirical evidence reported.
AI reconfigures ecosystems and platforms around foundation models, shifting how complementary actors interact and altering platform/ecosystem structure.
Analytical review of Ecosystems and Platforms literature; conceptual claim that foundation models act as central coordinating technologies; no empirical data or sample.
AI embeds algorithmic actors into the microfoundations of strategy, altering the role and behavior of individual-level actors that underlie firm-level phenomena.
Conceptual analysis of Microfoundations literature; theoretical proposition that algorithms act as actors at micro levels; no empirical sample provided.
AI creates hybrid cognitive architectures by integrating algorithmic cognition with human cognition, thereby changing how strategic decisions are made.
Theoretical argument drawing on literature in Behavioral Strategy and cognitive theory; conceptual synthesis without reported empirical tests or sample.
AI introduces a theoretical discontinuity that challenges core assumptions of strategic management (specifically those rooted in industry-structure and resource-based perspectives).
Conceptual/theoretical analysis across literatures in strategic management; the paper synthesizes prior debates and argues AI undermines prior assumptions. No empirical sample or quantitative data reported.
Some merged PRs introduce new lint or security findings while simultaneously removing existing issues (i.e., merges sometimes involve both addition and removal of issues).
Before-and-after static analysis (Pylint and Bandit) of merged PRs showing coexistence of introduced and removed findings in observed diffs.
We examine algorithmic co-supervision (ACoS) as a hybrid control mode in which supervisors and AC systems jointly direct, evaluate, and discipline workers.
The paper's stated empirical and conceptual focus; supported by the authors' analysis of 14 real-world ACoS settings (as reported in abstract).
Managerial authority is shifting from human supervisors alone toward varying hybrid arrangements in which humans and algorithms jointly control workers.
Claim drawn from prior literature and the authors' conceptual framing; the paper also analyzes real-world settings (14) to illustrate hybrid arrangements.
Cross-model validation reveals architecture-level trade-offs independent of specific LLMs: Dual Process excels at numeric/temporal queries (65-90% accuracy) while RAG excels at historical retrieval (60-85% accuracy).
Empirical cross-model tests across six LLMs; reported accuracy ranges for different query types and architectures.
Clarifying-question prompts produced mean rubric scores of 6.67 out of 8, higher than raw prompts but lower than checklist-improved prompts.
Reported mean rubric scores in the abstract showing clarifying-question prompts scored 6.67, compared to 5.67 for raw and 7.50 for checklist.
Classical categories (labour, capital, firm, market, productivity, trust) remain necessary but are incomplete for describing economic action when technologies prepare decisions, coordinate workflows, support tasks, verify transactions, and reshape responsibility.
Conceptual analysis supported by diagnostic indicators showing distributed decision/action capacity across humans, AI agents, robots, protocols, compute and energy systems; argumentative/theoretical evidence rather than causal inference.
Labour projections are more consistent with task reallocation than labour disappearance.
Analysis of labour-market reallocation data and labour projections (public sources) interpreted under a task-reallocation framework rather than full employment loss, using relative growth and reallocation indicators.
The central challenge is whether commercial influence in generative systems can be made trustworthy, i.e., attributable, measurable, contestable, and aligned with user welfare.
Normative claim and formulation of research and policy challenge presented by the authors as the central problem motivating the paper; based on their analysis of gaps in detection, measurement, and governance.
This reframes generative AI advertising as a problem of trustworthy intervention rather than content placement.
Authors' normative and conceptual reframing based on their analysis and taxonomy; presented as an argument about how to think about regulatory and design priorities.
High-AIC participants realized outsized gains from GenAI access; low-AIC participants saw limited or even negative marginal returns.
Subgroup analysis of the randomized experiment comparing treatment effects by AIC level; authors report large positive treatment effects for high-AIC subgroup and small or negative effects for low-AIC subgroup.
The distribution of gains from GenAI access was highly uneven across users.
Experimental results showing heterogeneous effects across participants (variance/heterogeneity analyses reported in the paper).
The system is generically bistable, with a stable partial adoption equilibrium coexisting alongside full genuine adoption.
Analytical results from the evolutionary game-theoretic model demonstrating multiple stable equilibria (bistability). No empirical sample (theoretical proof / model analysis).
Doctors choose among three strategies: genuine adoption, partial adoption, and rejection, where genuine adoption is required for systemic benefits to materialise above a population threshold.
Model specification in an evolutionary game-theoretic framework; analytical description of strategy set and threshold condition. No empirical sample (theoretical model).
The future of work will be shaped by decisions made at every level of society.
Normative/concluding statement in the chapter; presented as an implication of the prior analysis rather than an empirically tested claim.
AI affects the labour market through four channels: evolution of existing roles, creation of entirely new ones, redistribution across geographies and demographics, and selective displacement concentrated among older and lower-mobility workers.
Chapter synthesises labour market data, historical analogy, and emerging workplace evidence to propose these four channels; selective displacement claim references demographic concentration (older and lower-mobility workers).
Adaptation determines who benefits from technological (AI) change.
One of five lessons; argued using historical analogy and labour market patterns (qualitative claim in chapter).
LLMs often generate responses with the structural clarity associated with early-career engineers, yet they display persistent weaknesses in factual grounding and contextual interpretation.
Qualitative and comparative analysis of LLM responses against the expert rubric during the audit (six commercial LLMs); observed patterns in response form and substantive content.
There is a governance–task decoupling: under structural stress, text-only governance degrades on both governance and task dimensions simultaneously, whereas mechanical enforcement preserves governance quality even as task performance drops.
Experimental stress tests or structural-stress scenarios applied to both governance architectures in the paper's synthetic experiments; observed differential behavior across governance and task metrics. Abstract does not provide numeric details.
The improvement from mechanical enforcement is driven by architectural separation: LLM-generated rationales under mechanical enforcement show comparable CDL to text-only governance — the gain comes from removing clear-cut decisions from the model's control.
Analysis comparing LLM-generated rationales and a metric called CDL across governance architectures in the synthetic banking experiments; authors attribute improvement to removing certain decisions from the model's control. Specific statistics and CDL definition not provided in abstract.
Differences in human intervention effectiveness across escalation types are partly explained by variation in workers' post-escalation intervention effort.
Observed correlations (and subgroup comparisons) in the randomized experiment showing that measures of post-escalation effort (e.g., message counts, share of chat rounds, proactivity) vary across escalation types and relate to outcome differences.
Artificial intelligence (AI) is rapidly reshaping knowledge-intensive work by automating, augmenting, and reconfiguring core professional activities.
Paper asserts this as a motivating observation based on prior literature and descriptive claims; no original empirical sample or quantified data reported.
Metis can be subdivided into 'constitutive metis' (knowledge destroyed by the act of formalization) and 'operational metis' (system-specific familiarity that automation can progressively absorb).
Conceptual taxonomy proposed by the authors; definitions and distinctions are theoretical and illustrated via argumentation and prior literature rather than quantified empirical measurement.
Perceived procedural improvement (participants preferring facilitation and higher reported trust) can coexist with measurable steering of outcomes and unchanged participation inequality, motivating evaluation practices treating outcomes, interaction dynamics, and perceptions as distinct governance targets.
Synthesis of the experimental findings: null effect on consensus and participation equity, positive effects on participant preference/trust, and measurable allocation shifts (up to 5.5 percentage points) across facilitation conditions in the two experiments (total N=879).
Facilitators shifted select charity-level allocations by up to 5.5 percentage points, directly affecting the final charitable payout.
Analysis of final group allocation outcomes across experimental conditions showing shifts in allocation to specific charities; reported maximum observed shift of 5.5 percentage points attributable to facilitator condition(s). (Study-level sample covering the two experiments; participants organized in groups of three.)
Augmented work agency is shaped by whether applications are generative or non-generative, by employees' experiences of anxiety and technostress, and by micro-politics through which teams negotiate AI use and AI ethics.
Thematic findings from semistructured interviews (28 participants) and document review identifying these factors as shaping agency in practice.
The analysis uncovers three central tensions shaping AI-mediated work: autonomy versus orchestration; capability versus dependency; and experimentation versus ethics.
Recurring themes identified through qualitative interviews (28 participants) and document review; interpretive synthesis presented in findings.
AI integration transforms managerial practices, workforce identities and organizational coordination.
Thematic and interpretive analysis of semistructured interviews with 28 managers/professionals across 12 organizations and review of organizational documents.
Accounting for heterogeneity in AI literacy (agents' ability to identify and adapt to inaccurate AI outputs) can produce skill polarization in the long-run steady state.
Analytical/theoretical steady-state distribution analysis of agent skill dynamics with heterogeneous AI literacy parameters; paper reports conditions under which polarization emerges (theoretical, no empirical sample).
Beyond length biases, fine-tuning amplifies sycophancy and relationship-seeking behaviours in models.
Behavioral analysis of model outputs in the within-subject experiment (530 participants) showing increased incidence/intensity of sycophantic and relationship-seeking responses after preference fine-tuning compared to baseline models.
Adapting to individual preference data yields only marginal gains over training on pooled preferences from a diverse population.
Comparison within the same within-subject experiment (530 participants) between models fine-tuned on individual preferences versus models trained on pooled preferences across participants; reported as 'marginal gains'.
The dominant explanation for the gap locates it in model capability; instead, software-engineering capability emerges from a model-harness-environment system where a runtime substrate (the harness) mediates how an agent observes a project, acts on it, receives feedback, and establishes that a change is complete.
Conceptual argument and reframing presented in the paper (abstract). The paper formalizes this perspective rather than reporting a large-scale empirical test in the abstract.
There is a quality–motivation dissociation in AI-assisted goal-setting: AI-authored goals are objectively higher quality but produce lower motivation and worse behavioral follow-through.
Synthesis of experimental findings from the preregistered trial: higher SMART scores for LLM goals (d = 2.26) combined with lower self-reported motivation measures and lower two-week follow-up action rates.