Evidence (4793 claims)
Adoption
5539 claims
Productivity
4793 claims
Governance
4333 claims
Human-AI Collaboration
3326 claims
Labor Markets
2657 claims
Innovation
2510 claims
Org Design
2469 claims
Skills & Training
2017 claims
Inequality
1378 claims
Evidence Matrix
Claim counts by outcome category and direction of finding.
| Outcome | Positive | Negative | Mixed | Null | Total |
|---|---|---|---|---|---|
| Other | 402 | 112 | 67 | 480 | 1076 |
| Governance & Regulation | 402 | 192 | 122 | 62 | 790 |
| Research Productivity | 249 | 98 | 34 | 311 | 697 |
| Organizational Efficiency | 395 | 95 | 70 | 40 | 603 |
| Technology Adoption Rate | 321 | 126 | 73 | 39 | 564 |
| Firm Productivity | 306 | 39 | 70 | 12 | 432 |
| Output Quality | 256 | 66 | 25 | 28 | 375 |
| AI Safety & Ethics | 116 | 177 | 44 | 24 | 363 |
| Market Structure | 107 | 128 | 85 | 14 | 339 |
| Decision Quality | 177 | 76 | 38 | 20 | 315 |
| Fiscal & Macroeconomic | 89 | 58 | 33 | 22 | 209 |
| Employment Level | 77 | 34 | 80 | 9 | 202 |
| Skill Acquisition | 92 | 33 | 40 | 9 | 174 |
| Innovation Output | 120 | 12 | 23 | 12 | 168 |
| Firm Revenue | 98 | 34 | 22 | — | 154 |
| Consumer Welfare | 73 | 31 | 37 | 7 | 148 |
| Task Allocation | 84 | 16 | 33 | 7 | 140 |
| Inequality Measures | 25 | 77 | 32 | 5 | 139 |
| Regulatory Compliance | 54 | 63 | 13 | 3 | 133 |
| Error Rate | 44 | 51 | 6 | — | 101 |
| Task Completion Time | 88 | 5 | 4 | 3 | 100 |
| Training Effectiveness | 58 | 12 | 12 | 16 | 99 |
| Worker Satisfaction | 47 | 32 | 11 | 7 | 97 |
| Wages & Compensation | 53 | 15 | 20 | 5 | 93 |
| Team Performance | 47 | 12 | 15 | 7 | 82 |
| Automation Exposure | 24 | 22 | 9 | 6 | 62 |
| Job Displacement | 6 | 38 | 13 | — | 57 |
| Hiring & Recruitment | 41 | 4 | 6 | 3 | 54 |
| Developer Productivity | 34 | 4 | 3 | 1 | 42 |
| Social Protection | 22 | 10 | 6 | 2 | 40 |
| Creative Output | 16 | 7 | 5 | 1 | 29 |
| Labor Share of Income | 12 | 5 | 9 | — | 26 |
| Skill Obsolescence | 3 | 20 | 2 | — | 25 |
| Worker Turnover | 10 | 12 | — | 3 | 25 |
Productivity
Remove filter
The effect of AI adoption on widening the electricity output growth gap is more pronounced for firms located in economically advanced regions.
Heterogeneity analysis by regional economic development level using the firm-level electricity consumption dataset; stratified or interaction regressions showing larger estimated effects in more advanced regions. Exact subgroup sizes not provided in the summary.
The main result (initial widening of electricity growth gap) is robust to alternative variable definitions, exclusion of firms relying on outsourced AI services or non-AI adoption samples, and controls for endogeneity.
Robustness checks reported in the paper: alternative variable definitions, sample restrictions (excluding outsourced-AI-reliant firms and non-AI samples), and application of endogeneity control methods (e.g., instrumental variables or panel fixed effects). Exact methods and sample sizes not specified in the summary.
AI adoption initially widens the corporate electricity output growth gap at the firm level in China.
Empirical analysis using unique firm-level data on corporate electricity consumption in China; econometric estimation comparing electricity output growth between AI-adopting firms and non-adopting peers (panel/firm-level analysis). Sample size not stated in the summary.
To optimize agentic AI integration and ensure responsible innovation across financial services, interdisciplinary, longitudinal research and robust governance frameworks are needed.
Authors' conclusions and recommendations based on the identified findings and gaps in the reviewed literature.
Diverse architectural models such as multi-agent systems and cloud-based frameworks enable scalable, adaptive agentic AI deployments in financial services.
Synthesis of architecture-focused studies and framework descriptions within the reviewed literature (architectural benchmarking across papers).
Findings reveal substantial productivity gains and operational efficiencies predominantly in banking and investment.
Systematic review synthesizing multidisciplinary qualitative, quantitative, and bibliometric studies of agentic AI applications in financial services published up to mid-2024 (review-level synthesis).
The ManagerWorker two-agent pipeline (expensive text-only manager + cheaper worker with repo access) can substitute expensive execution by using expensive reasoning in the manager and cheaper execution in the worker.
System design description plus empirical results on 200 SWE-bench Lite instances showing parity in success rates between a strong-manager/weak-worker pipeline and a strong single agent while using fewer strong-model tokens.
A minimal review-only manager loop adds only 2 percentage points over the baseline, whereas structured exploration and planning by the manager add 11 percentage points, demonstrating that active direction (not mere reviewing) produces most of the benefit.
Ablation-style comparison of pipeline variants on the 200-instance SWE-bench Lite evaluation: review-only manager loop versus manager with structured exploration and planning; reported improvements in percentage points.
A strong manager directing a weak worker achieves a 62% success rate on software-engineering tasks, matching a strong single agent which achieves 60%, while using a fraction of the strong-model token usage.
Empirical evaluation on 200 instances from SWE-bench Lite across five pipeline configurations and model pairings; measured task success rates and token usage for manager-worker pipelines versus single-agent baselines.
Under economy-wide deployment, the share of computer-vision-exposed labor compensation that is cost-effectively automatable rises sharply (relative to the firm-level 11% estimate).
Model counterfactuals or calibration scenarios comparing firm-level deployment vs economy-wide deployment; qualitative statement that share increases substantially.
At the firm level, cost-effective automation captures approximately 11% of computer-vision-exposed labor compensation.
Calibration and implementation in computer vision; reported firm-level estimate from the framework.
Scale of deployment is a key determinant: AI-as-a-Service and AI agents spread fixed costs across users, sharply expanding economically viable tasks.
Modeling and calibration arguments showing fixed-cost spreading effects increase set of tasks for which automation is cost-effective; qualitative and quantitative comparisons in implementation.
Because higher accuracy is disproportionately costly (convex cost), full automation is often not cost-minimizing; partial automation, where firms retain human workers for residual tasks, frequently emerges as the equilibrium.
Theoretical model combined with calibration (scaling laws + task mappings); equilibrium outcomes reported from the framework implementation.
We model automation intensity as a continuous choice in which firms minimize costs by selecting an AI accuracy level, from no automation through partial human-AI collaboration to full automation.
The paper develops a theoretical framework / model that treats automation intensity as a continuous decision variable; described as the central modeling approach.
The findings demonstrate that technological innovation strategies, when effectively implemented, provide measurable competitive advantages for banks and offer evidence-based insights for policymakers and practitioners.
Authors' interpretation/conclusion drawing on the reported statistically significant relationships between innovation (product and technological) and competitiveness.
Technological innovation is positively and statistically significantly related to bank competitiveness (simple linear regression result reported).
Simple linear regression reported in the paper testing the hypothesis that technological innovation influences competitiveness; data collected from innovation-focused executives across licensed banks (paper states data from 39 licensed banks).
Product innovation strategy has a positive and statistically significant effect on competitiveness (F(1,134) = 74.983, p < .001).
Bivariate regression analysis reported in the paper with F(1,134)=74.983, p < .001; based on survey data from innovation-focused executives (regression degrees of freedom indicate n≈136 observations).
In the user study, AI-expanded 5W3H prompts increase user satisfaction from 3.16 to 4.04.
Reported pre/post or baseline vs AI-expanded satisfaction scores in the N=50 user study with numeric scores 3.16 and 4.04.
In the user study, AI-expanded 5W3H prompts reduce interaction rounds by 60 percent.
Reported comparison in the N=50 user study between baseline interaction rounds and rounds after AI-assisted 5W3H expansion; percentage reduction reported as 60%.
A weak-model compensation pattern was observed: the lowest-baseline model (Gemini) shows a much larger D-A gain (+1.006) than the strongest model (Claude, +0.217).
Model-level comparison of D-A gain (difference between structured and unstructured conditions) across three models (Claude, GPT-4o, Gemini) on the evaluated outputs; reported gains for Gemini and Claude.
The strongest structured conditions reduce cross-language sigma from 0.470 to about 0.020.
Reported numeric comparison of sigma (variance) between unstructured baseline and strongest structured prompting conditions across evaluated outputs.
Structured prompting substantially reduces cross-language score variance relative to unstructured baselines.
Empirical comparison across 3,240 outputs evaluated by DeepSeek-V3, comparing structured vs. unstructured prompting across three languages.
Prior work showed that PPS (Prompt Protocol Specification), a 5W3H-based structured intent framework, improves goal alignment in Chinese and generalizes to English and Japanese.
Statement referring to prior work (not new experiments in this paper); no sample size or methods provided in this text excerpt.
Together, these results suggest that ASI-Evolve represents a promising step toward enabling AI to accelerate AI across the foundational stages of development, offering early evidence for the feasibility of closed-loop AI research.
Aggregate of reported experimental results across architecture design, pretraining data curation, reinforcement learning algorithm design, and preliminary transfer experiments.
In reinforcement learning algorithm design, discovered algorithms outperform GRPO by up to +5.04 points on OlympiadBench.
Reinforcement learning algorithm design experiments reported in the paper comparing discovered algorithms to GRPO on OlympiadBench.
In reinforcement learning algorithm design, discovered algorithms outperform GRPO by up to +11.67 points on AIME24.
Reinforcement learning algorithm design experiments reported in the paper comparing discovered algorithms to GRPO on AIME24.
In reinforcement learning algorithm design, discovered algorithms outperform GRPO by up to +12.5 points on AMC32.
Reinforcement learning algorithm design experiments reported in the paper comparing discovered algorithms to GRPO on AMC32.
In pretraining data curation, gains exceed 18 points on MMLU.
Reported experimental result on MMLU benchmark within pretraining data curation experiments.
In pretraining data curation, the evolved pipeline improves average benchmark performance by +3.96 points.
Pretraining data curation experiments reported in the paper showing an average benchmark performance improvement of +3.96 points.
The best discovered model surpasses DeltaNet by +0.97 points, nearly 3x the gain of recent human-designed improvements.
Reported performance comparison between the best discovered model and DeltaNet in neural architecture experiments; statement comparing relative gain to recent human-designed improvements.
In neural architecture design, it discovered 105 SOTA linear attention architectures.
Neural architecture design experiments reported in the paper, with 105 discovered architectures labeled as SOTA.
ASI-Evolve augments standard evolutionary agents with two key components: a cognition base that injects accumulated human priors into each round of exploration, and a dedicated analyzer that distills complex experimental outcomes into reusable insights for future iterations.
Method description of ASI-Evolve's architecture/components in the paper (cognition base and analyzer added to evolutionary agents).
We present ASI-Evolve, an agentic framework for AI-for-AI research that closes this loop through a learn-design-experiment-analyze cycle.
Methodological contribution described in the paper: presentation and implementation of the ASI-Evolve framework and its learn-design-experiment-analyze loop.
Large language model (LLM) use can improve observable output and short-term task performance.
Paper synthesizes empirical findings from human–AI interaction studies, learning-research experiments, and model-evaluation work indicating improved produced outputs and short-term task performance when humans use LLMs; no single pooled sample size or unified effect estimate is reported in the paper.
These empirical insights provide actionable guidelines advocating dynamically routed architectures that adapt their collaborative structures to real-time task complexity.
Authors' recommendation derived from reported empirical findings comparing architectures under varying time budgets and task complexities (prescriptive claim based on study results).
Given extended compute budgets, the agent team topology achieves the deep theoretical alignment necessary for complex architectural refactoring.
Empirical benchmarks run with longer/extended computational budgets showing agent teams perform better on complex architectural refactoring tasks (qualitative claim; no numeric effect sizes or sample counts provided in the abstract).
The subagent mode functions as a highly resilient, high-throughput search engine optimal for broad, shallow optimizations under strict time constraints.
Benchmark comparisons in the execution-based testbed under strictly fixed computational time budgets showing subagent architecture excels in throughput/resilience for broad, shallow optimization tasks (qualitative claim in paper; no numeric effect sizes provided).
Autor et al. (2024) show that the majority of current employment is in job specialties that did not exist in 1940, with new task creation driven by augmentation-type innovations.
Citation reported in the paper summarizing Autor et al. (2024); no sample size provided in excerpt.
Firms may not sufficiently account for non-monetary aspects of technological progress (well-being, safety, quality of work); a planner would include such considerations in steering technological progress.
Normative conclusion based on theoretical analysis comparing firm objective functions (profits) vs social planner objectives (including non-monetary utility).
The planner can raise social welfare by focusing technological progress on making goods cheaper that are disproportionately consumed by relatively poorer agents, thereby raising their real income.
Extension of the baseline model to multiple goods showing distributional gains via composition of price changes (real income channel).
When capital and labor are gross complements, a planner concerned with workers' welfare would favor capital-augmenting innovations to raise wages.
Analytical result from the model analyzing factor-augmenting technological progress and complementarity between capital and labor.
A planner with sufficient welfare weight on workers will impose positive robot taxes, with the tax rate increasing in the planner's concern for workers' welfare.
Application of the baseline model to robot taxation; analytical derivation of optimal robot tax under planner preferences.
As labor's economic value diminishes, steering progress focuses increasingly on enhancing human well-being (non-monetary aspects) rather than labor productivity.
Theoretical discussion and model results in the paper showing planner's shifting objective when labor is devalued.
The welfare benefits of steering technology are greater the less efficient social safety nets are.
Analytical result from the paper's theoretical model comparing a planner who can/cannot perform transfers and evaluating steering as second-best when redistribution is costly.
These household-level non-market productivity gains (ChatGPT making productive online tasks more efficient and freeing time for leisure) are economically large and likely constitute a substantial share of the overall economic impact of generative AI.
Combination of empirical IV estimates showing leisure increases and productivity-unchanged productive time, plus model-implied efficiency gains; authors' interpretation and welfare discussion in paper.
Mapping the empirical time-reallocation into a quantitative household time-allocation model implies generative AI approximately doubles the efficiency of productive online tasks for adopters; preferred calibration implies efficiency gains of 76%–176%.
Quantitative time-allocation model adapted from Aguiar et al. (2021); model uses empirical IV estimates for time reallocation and Engel curve elasticities estimated via IV (local precipitation shocks). Authors report implied efficiency gains of 76%–176% and state 'approximately doubles' efficiency.
Households predominantly utilize ChatGPT in the context of productive online activities (education, job search, informational research) rather than during leisure browsing, as inferred from the browsing context around ChatGPT use.
High-frequency analysis comparing 30-minute browsing intervals around ChatGPT visits to intervals of demographically similar non-users; LLM-based inference of website purpose; observed co-occurrence with productive-site categories.
ChatGPT adoption increases the leisure share of browsing duration by about 30 percentage points.
IV long-difference estimates from Comscore browsing data with LLM-based site classification; authors report a ~30 percentage point increase in leisure share after adoption.
In long-difference IV estimates, ChatGPT adoption raises total leisure browsing time by roughly 150 log points.
IV long-difference estimates using pre-ChatGPT exposure as instrument; reported effect described as 'roughly 150 log points' increase in total leisure browsing time.
A household's pre-ChatGPT ex-ante exposure (based on 2021 browsing composition) strongly predicts subsequent ChatGPT adoption: a 1 SD higher exposure predicts a 2.5 percentage point higher rate of having used ChatGPT by December 2024.
Constructed 'exposure' measure by aggregating site-level overlap with chatbot capabilities over household 2021 browsing; predictive regression (household-level) linking 1 SD change in exposure to 2.5pp higher adoption by Dec 2024 (statistic reported in paper).