Evidence (3062 claims)
Adoption
5227 claims
Productivity
4503 claims
Governance
4100 claims
Human-AI Collaboration
3062 claims
Labor Markets
2480 claims
Innovation
2320 claims
Org Design
2305 claims
Skills & Training
1920 claims
Inequality
1311 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 373 | 105 | 59 | 439 | 984 |
| Governance & Regulation | 366 | 172 | 115 | 55 | 718 |
| Research Productivity | 237 | 95 | 34 | 294 | 664 |
| Organizational Efficiency | 364 | 82 | 62 | 34 | 545 |
| Technology Adoption Rate | 293 | 118 | 66 | 30 | 511 |
| Firm Productivity | 274 | 33 | 68 | 10 | 390 |
| AI Safety & Ethics | 117 | 178 | 44 | 24 | 365 |
| Output Quality | 231 | 61 | 23 | 25 | 340 |
| Market Structure | 107 | 123 | 85 | 14 | 334 |
| Decision Quality | 158 | 68 | 33 | 17 | 279 |
| Fiscal & Macroeconomic | 75 | 52 | 32 | 21 | 187 |
| Employment Level | 70 | 32 | 74 | 8 | 186 |
| Skill Acquisition | 88 | 31 | 38 | 9 | 166 |
| Firm Revenue | 96 | 34 | 22 | — | 152 |
| Innovation Output | 105 | 12 | 21 | 11 | 150 |
| Consumer Welfare | 68 | 29 | 35 | 7 | 139 |
| Regulatory Compliance | 52 | 61 | 13 | 3 | 129 |
| Inequality Measures | 24 | 68 | 31 | 4 | 127 |
| Task Allocation | 71 | 10 | 29 | 6 | 116 |
| Worker Satisfaction | 46 | 38 | 12 | 9 | 105 |
| Error Rate | 42 | 47 | 6 | — | 95 |
| Training Effectiveness | 55 | 12 | 11 | 16 | 94 |
| Task Completion Time | 76 | 5 | 4 | 2 | 87 |
| Wages & Compensation | 46 | 13 | 19 | 5 | 83 |
| Team Performance | 44 | 9 | 15 | 7 | 76 |
| Hiring & Recruitment | 39 | 4 | 6 | 3 | 52 |
| Automation Exposure | 18 | 16 | 9 | 5 | 48 |
| Job Displacement | 5 | 29 | 12 | — | 46 |
| Social Protection | 19 | 8 | 6 | 1 | 34 |
| Developer Productivity | 27 | 2 | 3 | 1 | 33 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
| Creative Output | 15 | 5 | 3 | 1 | 24 |
| Skill Obsolescence | 3 | 18 | 2 | — | 23 |
| Labor Share of Income | 8 | 4 | 9 | — | 21 |
Human Ai Collab
Remove filter
Contemporary AI (CNNs for imaging, LLMs for language) reliably automates narrowly defined clinical tasks and improves reproducibility and workflow efficiency, but cannot replace physicians in the foreseeable future.
Narrative literature review synthesizing empirical evaluations of convolutional neural networks in medical imaging and benchmarks/assessments of large language models; survey of studies reporting task-level accuracy, reproducibility, and workflow time-savings. Review is non-systematic (no meta-analysis).
Automation bias and changing work processes imply re‑skilling needs for public servants and potential shifts in public sector employment composition.
Findings and recommendations in multiple studies within the review documenting automation effects on workflows and workforce skill requirements (from the 103‑item corpus).
Predictive governance can change fiscal timing (earlier interventions) and alter uncertainty profiles for public budgets, requiring economists to model dynamic fiscal impacts and risks from algorithmic failure or bias.
Implication drawn in the review from case studies and economic reasoning present in the literature; recommendation for fiscal modeling based on synthesized evidence across the 103 items.
Interoperability and ethical‑by‑design requirements influence vendor lock‑in, competition, and the emergence of platform providers in markets for public‑sector AI solutions.
Policy and market analyses within the reviewed literature that link technical standards and ethical design requirements to market structure and vendor dynamics (synthesized from the 103 items).
Predictive analytics and AI enable anticipatory policy design (early intervention, forecasting), but they raise normative and governance questions about acceptable levels of prediction‑driven intervention.
Thematic findings from the review's mapping of predictive analytics use cases and accompanying ethical/governance discussions across the 103‑item corpus.
Human–AI interaction issues—such as automation bias and shifting public servant roles—affect decision quality and legitimacy, creating a need for human‑in‑the‑loop processes.
Multiple empirical and theoretical contributions in the reviewed literature identified automation bias and role shifts; recommendation for human‑in‑the‑loop emerges from synthesis of these studies.
Firms can realize productivity gains from adopting LLMs, but net value depends on verification, security remediation, and IP-management costs.
Firm-level case studies and productivity measurements in the literature showing time savings but also nontrivial verification/remediation effort; synthesis emphasizes net effect conditional on costs.
Adoption outcomes are shaped not only by technology and costs but also by customer perceptions, worker acceptance, and managerial actions; thus stakeholder-centered strategies are needed for successful deployment.
Synthesis of theoretical results from the evolutionary game (Essay 2) and the differentiated competition framework (Essay 1), supported by simulation experiments highlighting the role of perceptions and incentives. This is an interpreted conclusion rather than a direct empirical finding.
Adoption likelihood is sensitive to initial conditions and to parameters such as employee sensitivity to robots, training costs, perceived risks, marketing influence, and labor efficiency.
Sensitivity analysis within the MATLAB simulations varying parameters (training costs, perceived risk, marketing strength, labor efficiency) and initial states; evolutionary game theoretical structure showing path dependence.
Stakeholder attitudes toward AI service robots evolve strategically; widespread positive adoption requires favorable initial conditions and appropriate incentives (e.g., lower training costs, higher labor efficiency, effective marketing).
Analytical framework: three-player evolutionary game theory modeling hotel owners, employees, and customers; computational evidence from MATLAB simulations and sensitivity analysis that vary initial states and parameters (training costs, perceived risks, marketing strength, labor efficiency) to map dynamic trajectories and basins of attraction.
AI adoption has a U-shaped effect on hospitality firm profit: short-term costs and adjustment can reduce profits, while longer-term gains from differentiation and productivity raise profits.
Combined theoretical analysis (differentiated Bertrand competition model incorporating demand-side differentiation and productivity mechanisms) and an empirical firm-level analysis reported in the dissertation that links AI adoption measures to profit, demand, and productivity indicators. (Sample size and specific datasets not reported in the provided summary.)
Adoption frictions—integration costs, data access, reliability, and regulatory compliance—may slow diffusion of AI agents and create heterogeneity in economic value across firms and sectors.
Theoretical implication supported by observed orchestration and governance challenges in deployments; recommendation/interpretation rather than direct causal measurement.
Implementation heterogeneity (how guardrails, human oversight, and orchestration are configured) likely drives outcome variation across deployments.
Observed heterogeneity in Alfred AI deployments and stated limitation that configuration differences affect outcomes; based on deployment comparisons and qualitative analysis (sample size/configurations unspecified).
Net productivity gains may be smaller once indirect costs—governance, monitoring, error-correction, orchestration—are accounted for; standard productivity accounting should include these costs.
Conceptual argument supported by observational documentation of governance and monitoring burdens in deployments; no precise cost accounting reported in summary.
Autonomous agents are likely to substitute for routine, structured cognitive tasks while complementing higher-level managerial and strategic tasks, accelerating task reallocation within firms.
Synthesis of prior literature (generative AI productivity findings) and observational deployment patterns from Alfred AI indicating substitution of routine tasks and continued human involvement in oversight/strategy.
Realized productivity gains from AI agents are materially constrained by governance complexity, model reliability limits (errors, hallucinations, edge cases), orchestration challenges across tools/data/human teams, and continued need for human-in-the-loop oversight.
Qualitative operational impacts and deployment observations from Alfred AI implementations, documented frictions in policies, safety constraints, error handling, and orchestration; evidence drawn from observational deployments and operational logs.
Effectiveness and safety of AI agents require structured guardrails and human-in-the-loop designs; AI agents function as scalable cognitive infrastructure only conditional on such governance.
Synthesis of deployment experience and analysis of constraints; recommendation grounded in observed model reliability issues, governance complexity, and oversight needs from the Alfred AI experiments.
Deployment of AI agents shifts demand toward roles focused on oversight, orchestration, prompt/agent engineering, and governance, creating new types of labor that may offset some direct labor reductions.
Authors' inference based on observed need for human oversight and orchestration in deployments; not quantitatively measured in the study (no headcount or labor-share data reported).
AI‑enabled risk assessment (weather, pests, price forecasts) can improve index insurance and credit scoring for smallholders, lowering financing costs and increasing investment — but it also raises concerns about data bias and exclusion.
Pilot programs and modeling studies on index insurance and credit scoring, combined with policy analyses documenting equity and bias risks; primary empirical work is limited to pilots and simulations.
Returns to AI investments depend on complementary investments in farmer knowledge, extension services, and local institutions; AI tends to amplify returns to managerial skills and digital literacy.
Empirical studies and randomized/quasi‑experimental trials showing complementarity effects, and qualitative evidence from stakeholder interviews; cited studies report larger impacts where complementary services exist.
Impacts of technology‑ecology integration are heterogeneous: they vary by farm size, crop type, local infrastructure, and farmer skills; smallholders can benefit substantially but are more constrained by liquidity, information, and market access.
Observational econometric analyses and randomized/quasi‑experimental studies reporting heterogeneous treatment effects, supplemented by qualitative interviews and case studies documenting constraints faced by smallholders.
Labor-market consequences will involve reallocation effects: routine-task automation, rising returns to managerial and technical skills, and potential within-firm wage dispersion.
Synthesis of labor economics theory and prior empirical work on automation; book recommends matched employer-employee panel studies to trace these effects but does not report such new panel results.
AI’s effects vary by industry, task composition, and firm capabilities; high-data, standardized-task sectors see faster, deeper impacts.
Cross-sector examples and theoretical arguments about task routineness and data intensity; calls for heterogeneity-aware empirical designs (e.g., difference-in-differences with staggered adoption).
Automation of routine tasks raises demand for cognitive, interpersonal, and technical skills; firms face reskilling needs and changing task allocation between humans and machines.
Task-level analytic framework and literature review on automation effects; book recommends empirical approaches (e.g., occupation and job-task data) to quantify these changes but does not present a single large empirical estimate.
Managers shift from routine decision execution to tasks involving oversight, interpretation, strategic design, and ethical stewardship of AI systems.
Qualitative case studies and literature review of task-level research; suggested task-analytic methods rather than reporting a specific empirical task dataset.
AI complements some researcher tasks (idea generation, analysis, writing) and substitutes others (routine editing, literature searches), changing skill demand and training priorities.
Stated under Labor Market Effects. Supported conceptually and likely by task-level studies or surveys; abstract doesn't cite specific empirical evidence or measurement details.
Impacts of AI adoption are broad, affecting individual researcher productivity, team workflows, and institutional outcomes in scholarly communication and digital scholarship.
Key Points summary. Basis likely includes mixed-methods evidence (surveys/interviews at individual and team levels, case studies, platform usage data) synthesized in the paper; abstract lacks detail on scope and samples.
Investment in governance and training is a necessary cost to realize sustained returns from generative AI; these costs influence adoption timing and the distribution of benefits.
Conceptual argument from the review supported by case examples and economic reasoning about complementary investments.
There is a risk of wage polarization: increased returns to AI‑complementary skills and potential downward pressure on wages for automatable tasks.
Theoretical synthesis drawing on economic models of skill‑biased technological change and early empirical observations; no definitive causal wage studies reported.
Generative AI will drive occupational reallocation by substituting routine cognitive tasks while complementing higher‑order cognitive and monitoring skills.
Theoretical labor economics arguments synthesized with early empirical examples; no large‑scale causal labor market study provided in the review.
Routine, boilerplate, and debugging tasks are most automatable or complemented by LLMs, shifting value toward design, verification, and systems thinking.
Task-level analyses, observational studies, and synthesized findings showing larger gains on repetitive or templated tasks versus high-level design tasks.
Liability and intellectual-property ownership around AI-assisted code are unresolved practical and legal concerns.
Legal and policy analyses, practitioner reports, and qualitative interviews noting ambiguous legal frameworks and unresolved questions about ownership and liability for AI-assisted code.
The framework quantitatively captures trade-offs between public-health outcomes and economic stability across macroscopic scenarios and different LLM backends.
Quantitative analysis reported across scenarios and model variants, tracking trade-off metrics between health (infection curves) and economic outcomes (aggregate activity). The summary notes cross-backend comparisons but does not report numerical effect sizes.
When coupled with an epidemic–economic model, the LLM-PDA framework robustly generates divergent macro trajectories across scenarios.
Coupled epidemic and economic modules in simulation; experiments run across diverse macroscopic scenarios (varying transmissibility, shocks, policy regimes) with metrics tracked at macro scale (infection prevalence over time, aggregate economic indicators). (Number of scenarios and runs not specified in summary.)
TGAIF clarifies where GenAI acts as a complement (augmenting consultant capability) versus where it risks substitution.
Conceptual distinction and mapping presented in the TGAIF derived from practitioner accounts; theoretical/qualitative, not empirically quantified across tasks.
TGAIF implies reallocation of work away from GenAI‑suitable subtasks (routine synthesis, drafting, summarization) toward tasks where human judgment and client interaction add most value.
Based on authors' inductive analysis of practitioner interviews describing which subtasks firms consider suitable for GenAI and which require human oversight; qualitative, not quantitatively tracked reallocation.
Aligning consulting tasks with generative-AI capabilities via a Task–GenAI Fit (TGAIF) framework can unlock substantial efficiency gains while containing key risks (notably hallucinations and loss of skill retention).
Inductive framework developed from qualitative, interpretive interviews with practitioners at leading German management‑consulting firms. The abstract does not report sample size, interview protocol, or quantitative validation; evidence is based on practitioner reports and the authors' synthesis.
DAR implies changes to labor and contracting: reversible AI leadership reshapes task boundaries, demand for oversight skills, and should be reflected in contracts and procurement with explicit authority-reversal rules and audit obligations.
Theoretical/ normative argument in implications section; no empirical labor or contract data included.
AI substitutes for routine coding tasks but complements higher-order tasks such as system architecture, integration, and orchestration.
Interpretation from qualitative evidence at Netlight where practitioners used AI for routine chores while retaining control of higher-order design tasks; no quantitative task-time displacement data presented.
Human roles are shifting toward oversight, curation, specification, and orchestration of multiple AI components and tools.
Synthesized from practitioner descriptions and changing task allocations observed in the Netlight fieldwork (interviews/observations); no longitudinal measurement of role changes reported.
Short-run consumer gains from faster, cheaper service can be undermined by trust losses from hallucinations or perceived deception, reducing long-term consumer surplus.
Conceptual welfare analysis and cited case examples in the literature; no longitudinal consumer-surplus measurement provided in this review.
Conventional productivity metrics (e.g., handle time) may misstate value because they do not capture multi-dimensional impacts like quality and trust.
Conceptual critique and synthesis of measurement challenges discussed in the literature; no empirical measurement study presented in this review.
There is potential for substantial cost savings and throughput gains in repetitive, high-volume interactions, but these are offset by costs for integration, monitoring, and error remediation.
Industry case examples and conceptual cost–benefit reasoning aggregated in the review; the paper contains no new quantitative cost estimates or sample-based measurements.
Generative AI will substitute for routine service tasks while complementing skilled workers for escalations and complex problem solving, shifting labor demand toward supervisory and relationship-focused roles.
Economic and labor-market analyses synthesized in the review; projections are inferential and based on heterogeneous secondary sources, not primary labor-market experiments.
Full automation of customer service is suboptimal because persistent risks (hallucinations, contextual errors, lack of genuine empathy, integration complexity) remain; hybrid human–AI systems achieve the best outcomes.
Synthesis of documented failure modes and practitioner case examples from the literature; no primary experimental data or controlled trials in this review. Inference based on heterogeneous empirical reports and conceptual analyses.
Welfare effects of democratized access to AI-assisted ideation are ambiguous: access could democratize innovation but also amplify low-quality outputs and misinformation absent proper curation.
Theoretical discussion and empirical examples of misinformation/low-quality outputs from LLMs cited in the review; no comprehensive welfare accounting provided.
Net gains in innovation from increased idea volume depend on complementary human capacity for curation and development; raw increases in ideas do not automatically translate into higher-quality innovation.
Synthesis noting studies where idea quantity rose but downstream quality or successful development did not necessarily increase; review highlights heterogeneity across workflows and dependence on human integration.
The most effective deployment model is a 'cognitive co-pilot' in which AI expands and challenges the idea space while humans provide curation, strategic evaluation, and experiential judgment.
Prescriptive conclusion drawn from synthesis of studies where human-AI collaboration (human curation/selection) produced better downstream outcomes than AI-alone outputs; evidence heterogenous and largely short-term.
Generative AI functions as a dual-purpose cognitive tool: a high-volume catalyst for divergent idea generation and a structured assistant for decomposing complex problems.
Nano-review / synthesis of existing empirical literature on LLM-assisted creativity and problem-solving, drawing on experimental ideation tasks, design/ideation studies, and applied case evidence; no original dataset or new experiments in this paper.
Net value from generative AI is contingent: gains are largest where breadth of ideas and rapid iteration matter, and smaller or riskier where deep domain expertise, tacit knowledge, or high-stakes judgments are required.
Synthesis of heterogeneous empirical results showing task-dependent benefits; argument grounded in observed differences across lab and field contexts and documented limitations in domain-specific performance.