Evidence (2340 claims)
Adoption
5267 claims
Productivity
4560 claims
Governance
4137 claims
Human-AI Collaboration
3103 claims
Labor Markets
2506 claims
Innovation
2354 claims
Org Design
2340 claims
Skills & Training
1945 claims
Inequality
1322 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 378 | 106 | 59 | 455 | 1007 |
| Governance & Regulation | 379 | 176 | 116 | 58 | 739 |
| Research Productivity | 240 | 96 | 34 | 294 | 668 |
| Organizational Efficiency | 370 | 82 | 63 | 35 | 553 |
| Technology Adoption Rate | 296 | 118 | 66 | 29 | 513 |
| Firm Productivity | 277 | 34 | 68 | 10 | 394 |
| AI Safety & Ethics | 117 | 177 | 44 | 24 | 364 |
| Output Quality | 244 | 61 | 23 | 26 | 354 |
| Market Structure | 107 | 123 | 85 | 14 | 334 |
| Decision Quality | 168 | 74 | 37 | 19 | 301 |
| Fiscal & Macroeconomic | 75 | 52 | 32 | 21 | 187 |
| Employment Level | 70 | 32 | 74 | 8 | 186 |
| Skill Acquisition | 89 | 32 | 39 | 9 | 169 |
| Firm Revenue | 96 | 34 | 22 | — | 152 |
| Innovation Output | 106 | 12 | 21 | 11 | 151 |
| Consumer Welfare | 70 | 30 | 37 | 7 | 144 |
| Regulatory Compliance | 52 | 61 | 13 | 3 | 129 |
| Inequality Measures | 24 | 68 | 31 | 4 | 127 |
| Task Allocation | 75 | 11 | 29 | 6 | 121 |
| Training Effectiveness | 55 | 12 | 12 | 16 | 96 |
| Error Rate | 42 | 48 | 6 | — | 96 |
| Worker Satisfaction | 45 | 32 | 11 | 6 | 94 |
| Task Completion Time | 78 | 5 | 4 | 2 | 89 |
| Wages & Compensation | 46 | 13 | 19 | 5 | 83 |
| Team Performance | 44 | 9 | 15 | 7 | 76 |
| Hiring & Recruitment | 39 | 4 | 6 | 3 | 52 |
| Automation Exposure | 18 | 17 | 9 | 5 | 50 |
| Job Displacement | 5 | 31 | 12 | — | 48 |
| Social Protection | 21 | 10 | 6 | 2 | 39 |
| Developer Productivity | 29 | 3 | 3 | 1 | 36 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
| Skill Obsolescence | 3 | 19 | 2 | — | 24 |
| Creative Output | 15 | 5 | 3 | 1 | 24 |
| Labor Share of Income | 10 | 4 | 9 | — | 23 |
Org Design
Remove filter
The paper presents Governed Memory, a shared memory and governance layer addressing the memory governance gap.
System architecture and design description in the paper (proposal of a shared memory and governance layer).
A hybrid strategic–computational framework, supported by governance mechanisms (human-in-the-loop checkpoints, escalation paths, accountability structures), is motivated to manage tensions and ensure responsible decision-making in AI-rich managerial contexts.
Synthesis-driven prescriptive framework produced by cross-framework analysis; conceptual recommendation rather than implementation evidence.
Roles oriented to information processing, optimisation, and operational precision (monitor, disseminator, resource allocator) are substantially enhanced by computational thinking (automation, optimisation, algorithmic decision-support).
Theoretical mapping of computational capabilities onto Mintzberg’s information-processing roles; conceptual reasoning without empirical validation.
Investment choices in collaboration AI and digital infrastructure become central strategic decisions affecting firms' comparative advantage.
Management literature synthesis and illustrative multinational cases; argument is conceptual without firm‑level comparative empirical data presented in the paper.
AI collaboration tools (virtual assistants, meeting summarizers, asynchronous platforms) complement hybrid work by reducing coordination costs and supporting dispersed teamwork.
Conceptual integration of technology and organizational literature; supported by illustrative case examples of multinational organizations but not by new quantitative causal evidence.
Hybrid and remote work increase employee autonomy and work–life integration.
Conceptual synthesis of sociological and management literatures; supported by secondary data and illustrative case studies from multinational organizations. No primary quantitative analysis or sample size reported—based on comparative case illustrations and theoretical integration.
Generative AI functions as a socio‑technical intermediary that facilitates interpretation, coordination, and decision support rather than merely automating discrete tasks.
Thematic analysis and co‑word linkage between terms related to interpretative work, coordination, and decision‑support and technical GenAI terms within the corpus.
The literature indicates a managerial shift away from hierarchical command‑and‑control toward guide‑and‑collaborate paradigms, where managers curate, guide, and coordinate AI‑augmented teams rather than micro‑manage tasks.
Synthesis of themes from the 212‑paper corpus (co‑word and thematic analyses) showing recurrent managerial/behavioural concepts such as autonomy, coordination, and decision‑support tied to GenAI discussions.
Trust is a principal demand driver for AI-enabled marketing among Generation Z — higher trust substantially raises adoption intention and thereby accelerates diffusion.
Interpretation/implication drawn from the large standardized path coefficients (Trust → Adoption Intention β = 0.718) and mediation results in the SEM on n = 450 Gen Z respondents.
Adoption intention partially mediates the relationship between trust and brand loyalty (indirect effect Trust → Adoption → Loyalty: standardized β ≈ 0.390, p = 0.001).
Cross-sectional survey (n = 450); mediation tested within SEM framework; reported indirect standardized effect ≈ 0.390 with p = 0.001.
Economic models of firm behavior and market microstructure should incorporate endogenous, adaptive segmentation processes and faster feedback loops enabled by human–AI systems; ABS and large‑scale interaction data can be used to calibrate such models.
Methodological recommendation grounded in the study's mixed‑methods findings (ABS experiments and 150M interaction dataset) and observed differences between autopoietic and traditional STP regimes.
Canvas Design Principles mitigate algorithmic myopia (overfitting to historical patterns) and improve adaptability and resource efficiency.
Set of design principles proposed in the paper and evaluated through agent‑based simulation scenarios and analyses of the large behavioral dataset. Specific experimental details and quantitative effect sizes for these principles are not detailed in the summary.
Reconceptualizing STP as an autopoietic (self‑organizing) system enables continuous human–AI co‑creation and yields better outcomes in unstable markets than traditional, process‑based STP.
Conceptual argument grounded in 6‑month lab ethnography (n = 23), design and deployment of the Algorithmic Canvas in that lab context, and validation via large behavioral dataset analyses and agent‑based simulations.
Algorithmic co‑creation methods detect substantial market fluctuations about 5.8× better than traditional approaches.
Computational analysis of large behavioral dataset (150 million customer interactions) and comparative performance evaluation in empirically grounded agent‑based simulations. The detection metric and statistical significance details are not provided in the summary.
The autopoietic model shortens strategic planning cycle length by approximately 90%.
Observed/recorded time‑to‑update or strategy revision metrics gathered via Algorithmic Canvas usage and lab ethnography (6‑month lab ethnography inside a Fortune 500 company, n = 23). Exact measurement protocol and whether reduction measured in live firms, simulations, or system logs is not fully detailed in the summary.
Design and policy interventions that encourage active human contributions (e.g., draft-first workflows, co-creation interfaces, training) can help preserve worker agency and mitigate psychological costs.
Recommendation based on experimental evidence that Active-collaboration preserved psychological outcomes relative to passive use; presented as policy/design prescription rather than directly tested intervention at scale.
A complementary real-world survey (N = 270) across diverse tasks reproduced the experimental pattern, suggesting external validity beyond the lab writing tasks.
Cross-sectional survey of N = 270 respondents reporting on their AI use across multiple task types; reported patterns consistent with the experiment (passive use associated with lower efficacy/ownership/meaningfulness; active collaborative use did not).
Effective teams tend to evolve from ad-hoc interpretive methods toward systematic evaluation by (a) formalizing prompts/tests, (b) instrumenting outputs, (c) mapping failure modes to remediation paths, and (d) creating organizational decision rules.
Pattern observed in the qualitative coding of interviews where participants described trajectories or steps their teams took to formalize evaluation.
Successful teams close the results-actionability gap by systematizing interpretive practices and creating clearer pathways from evaluation signals to product changes.
Interview accounts and cross-case analysis showing some teams adopting formalization steps (e.g., standardized prompts/tests, instrumentation, remediation mappings) that participants described as enabling action.
Immediate research priorities for AI economists include: field experiments testing NLP‑driven acquisition/personalization (measuring CAC, LTV, retention, consumer welfare); structural/empirical models of adoption that include data access costs and complementarities; and analyses of privacy regulation impacts on external text data availability and value.
Authors' set of recommended research directions derived from identified gaps in the systematic review and implications for AI economics.
Unit costs for bookkeeping and compliance tasks are likely to fall, potentially affecting professional services pricing and leading to consolidation.
Analytic inference from case advantages and industry literature; no empirical market-wide cost study included.
Generative AI can raise labor productivity in finance and tax, shifting work from routine processing to oversight, exceptions handling, and higher-value analysis.
Analytical framing supported by case observations and literature; presented as an expected economic effect rather than measured across a population.
Successful deployment requires new human capital: finance professionals with AI literacy, data governance, model validation, and control expertise.
Paper's labor and skills implications derived from case examples and analytic framing; recommendation-based observation rather than measured workforce data.
Generative AI provided better decision support via scenario analysis and anomaly prioritization.
Descriptive case examples and literature indicating use of LLMs and RAG systems for drafting scenarios and prioritizing anomalies; evidence is qualitative and illustrative.
Generative AI adoption produced cost savings through labor reallocation and task automation.
Qualitative evidence from Xiaomi and Deloitte case analysis and analytic framing suggesting lower labor requirements for routine tasks; no standardized cost-accounting or sample-wide cost metrics provided.
Using generative AI led to higher consistency and reduced human error in repetitive finance/tax tasks.
Case-driven qualitative observations from the two organizational examples and literature synthesis indicating reduced variability in repetitive processes when AI-assisted.
Generative AI deployment increased processing speed and throughput for routine finance and tax tasks.
Observed improvements reported in case studies (Xiaomi and Deloitte) and corroborating industry/literature sources described in the paper; qualitative descriptions rather than standardized time-motion metrics.
Applying generative AI within corporate financial sharing centers (illustrated by Xiaomi’s Financial Sharing Center) and professional services firms (Deloitte) materially improves the efficiency and accuracy of finance and tax operations.
Qualitative case analysis of two organizations (Xiaomi Financial Sharing Center and Deloitte) supplemented by literature review and analytical mapping; no large-scale quantitative measurement reported.
Phased deployment and regulatory sandboxes can lower barriers for startups to pilot lower-risk applications, thereby shaping innovation trajectories.
Comparative policy analysis of sandboxing and phased deployment approaches in other jurisdictions; prescriptive inference without empirical testing in Vietnam.
Properly governed AI can yield large efficiency gains (reduced processing time and lower per-case costs), but those gains depend on redesigning legal processes to accommodate algorithmic workflows.
Analytic synthesis of administrative-process characteristics and AI capabilities; no primary quantitative evidence or measured effect sizes provided.
Establishing a graduated implementation model and clear regulatory pathways reduces regulatory uncertainty and makes public-sector AI procurement and private-market participation more predictable and attractive.
Normative recommendation informed by comparative institutional analysis and economic reasoning; not empirically tested in the paper.
A graduated implementation model—phased deployment, differentiated safeguards by risk, and mandatory human oversight for high-stakes decisions—can balance innovation with rule-of-law protections.
Normative framework development combining doctrinal findings and comparative lessons; prescriptive recommendation rather than empirical validation.
Comparative analysis of international frameworks reveals a range of institutional responses and regulatory instruments that Vietnam could adapt.
Comparative institutional analysis synthesizing governance approaches from liberal and civil-law jurisdictions (review of secondary sources and policy frameworks).
AI can substantially modernize administrative decision-making in civil-law systems (speed, consistency, scalability).
Qualitative doctrinal and comparative institutional analysis using Vietnam as a focused case study; no primary quantitative field data or sample size.
SERF (Structured Error Recovery Framework) defines structured, machine-readable failure semantics to enable deterministic agent self-correction and automated recovery strategies.
Design and formal specification of SERF in the paper; formalized as a testable hypothesis with reproducible experimental methodology.
ATBA (Adaptive Timeout Budget Allocation) frames sequential tool invocation as a budget-allocation problem over heterogeneous latency distributions to improve end-to-end latency and reliability.
Algorithmic formulation and formalization provided in the paper; ATBA is presented as a testable hypothesis with reproducible benchmarks and latency/error models.
The MCP (Model Context Protocol) is widely adopted: >10,000 active MCP servers and 97 million monthly SDK downloads as of early 2026.
Reported protocol-adoption metrics in the paper (protocol adoption context); presumably aggregated server and SDK-download statistics (time-stamped to early 2026).
Historical institutional publication records encode an extractable evaluative signal ("taste") that can be learned by models and used for scalable triage, screening, and curation of submissions.
Empirical results showing improved predictive accuracy after fine-tuning on accept/reject records, plus demonstration of transfer tasks and a cross-field (economics) result; implications for applications (triage, screening) are drawn from these empirical findings rather than directly deployed field experiments.
Models show well-calibrated confidence: their highest-confidence predictions are 100% accurate.
Calibration analysis of fine-tuned models comparing predicted-confidence levels to actual accuracy; reported that examples the model assigned its highest confidence to were 100% accurate. (Number of highest-confidence examples and calibration buckets not reported in the provided text.)
The learned evaluative signal transfers to untrained tasks such as pairwise comparisons and one-sentence summaries.
Fine-tuned models were evaluated on related, untrained evaluative tasks (pairwise comparisons of pitches and one-sentence summary evaluations) and showed positive transfer performance relative to baselines. (Specific metrics, effect sizes, and sample sizes for these transfer tasks are not provided in the supplied text.)
The core findings (harm from ToM order mismatches and benefits from A-ToM) are robust to partners beyond LLM-driven agents.
Paper reports robustness checks testing generalization to non-LLM agent classes (details summarized in robustness section); comparisons use the same coordination metrics.
A-ToM recovers coordination performance by aligning its effective ToM depth with partners across a range of multiagent tasks.
Experimental results showing A-ToM achieves coordination levels closer to matched fixed-order pairings across the repeated matrix game, grid navigation tasks, and Overcooked when facing partners with different fixed ToM depths.
An adaptive ToM (A-ToM) agent that infers its partner's ToM order from prior interactions and conditions its predictions and actions on that estimate restores alignment and improves coordination.
Implemented A-ToM (estimation from interaction history + conditioning of partner-action predictions) and evaluated it against fixed-order agents in the four environments; reported improvements in coordination metrics when A-ToM paired with partners of varying ToM orders.
Human–AI chat logs contain more explicit strategy commitments (stated rules) than human–human chats.
Content analysis / coding of natural-language chat logs from the human–AI experiment (human–AI n = 126) and the human–human benchmark (n = 108); coding counts show higher frequency of explicit commitments/statements of rules in human–AI messages.
Human–human subjects converge to Tit‑for‑Tat under one condition and to unconditional cooperation under the repeated-communication condition.
Strategy-estimation and behavioral trajectory analysis from the human–human benchmark (Dvorak & Fehrler 2024; n = 108) reported in the paper, showing condition-dependent convergence to Tit‑for‑Tat and to unconditional cooperation under repeated communication.
Strategy estimation indicates human–AI subjects tend to favor Grim Trigger when allowed pre-play communication.
Strategy-estimation/classification applied to subjects' choices in the human–AI condition with pre-play chat (subset of the human–AI n = 126); inferred strategy prevalence shows elevated assignment to Grim Trigger-type rules.
Extensive simulation experiments across different network topologies and attacker/defense scenarios validate both the FJ modeling of LLM-MAS and the effectiveness of the trust-adaptive defense.
Multiple simulation studies reported in the paper that vary network density, trust matrices, attacker stubbornness/persuasiveness, and defense strategies; validation claims stem from consistent patterns observed across these simulated settings. (The summary does not list the number of experimental runs or statistical reporting.)
A trust-adaptive defense that dynamically reduces trust in agents suspected of adversarial behavior can limit adversarial influence while preserving cooperative performance better than static trust-lowering strategies.
Implemented a trust-adaptive mechanism and evaluated it in simulation experiments across multiple network topologies and attack/defense scenarios, reporting reductions in adversarial sway with preserved task performance compared to naïve trust reduction. (Exact experimental counts and numeric effect sizes not provided in the summary.)
Increasing the number of benign agents dilutes an adversary's relative influence and thereby reduces the probability and magnitude of persuasion cascades.
Simulation experiments varying the count of benign agents in networks while measuring adversarial sway and collective opinion outcomes across different topologies. (Summary does not report exact counts or statistical summaries.)
The Friedkin–Johnsen opinion-dynamics model (innate opinions + interpersonal influence weights + stubbornness) closely captures LLM-MAS behavior across settings, both theoretically and empirically.
Modeling: derivation of FJ dynamics for LLM-MAS; Empirical: simulation experiments comparing FJ model predictions to observed LLM-MAS opinion trajectories and final consensus under varied topologies and trust matrices. (Exact goodness-of-fit metrics and sample counts not provided in the summary.)