The Commonplace
Home Dashboard Papers Evidence Syntheses Digests 🎲

Evidence (7395 claims)

Adoption
7395 claims
Productivity
6507 claims
Governance
5877 claims
Human-AI Collaboration
5157 claims
Innovation
3492 claims
Org Design
3470 claims
Labor Markets
3224 claims
Skills & Training
2608 claims
Inequality
1835 claims

Evidence Matrix

Claim counts by outcome category and direction of finding.

Outcome Positive Negative Mixed Null Total
Other 609 159 77 736 1615
Governance & Regulation 664 329 160 99 1273
Organizational Efficiency 624 143 105 70 949
Technology Adoption Rate 502 176 98 78 861
Research Productivity 348 109 48 322 836
Output Quality 391 120 44 40 595
Firm Productivity 385 46 85 17 539
Decision Quality 275 143 62 34 521
AI Safety & Ethics 183 241 59 30 517
Market Structure 152 154 109 20 440
Task Allocation 158 50 56 26 295
Innovation Output 178 23 38 17 257
Skill Acquisition 137 52 50 13 252
Fiscal & Macroeconomic 120 64 38 23 252
Employment Level 93 46 96 12 249
Firm Revenue 130 43 26 3 202
Consumer Welfare 99 51 40 11 201
Inequality Measures 36 105 40 6 187
Task Completion Time 134 18 6 5 163
Worker Satisfaction 79 54 16 11 160
Error Rate 64 78 8 1 151
Regulatory Compliance 69 64 14 3 150
Training Effectiveness 81 15 13 18 129
Wages & Compensation 70 25 22 6 123
Team Performance 74 16 21 9 121
Automation Exposure 41 48 19 9 120
Job Displacement 11 71 16 1 99
Developer Productivity 71 14 9 3 98
Hiring & Recruitment 49 7 8 3 67
Social Protection 26 14 8 2 50
Creative Output 26 14 6 2 49
Skill Obsolescence 5 37 5 1 48
Labor Share of Income 12 13 12 37
Worker Turnover 11 12 3 26
Industry 1 1
Clear
Adoption Remove filter
Operationalizing hardware-based governance must address transition realities including legacy hardware, attestation at scale, and protection of civil liberties.
Policy implementation analysis in the paper identifying practical challenges to deploying hardware-layer controls (conceptual/operational analysis; no empirical trial data provided).
high mixed The Open-Weight Paradox: Why Restricting Access to AI Models... practical hurdles to governance deployment (legacy hardware, attestation scalabi...
For LLM agents, memory management critically impacts efficiency, quality, and security.
Statement in paper framing and motivation; supported conceptually by literature linking memory design to system properties (no specific experimental details provided in abstract).
high mixed FSFM: A Biologically-Inspired Framework for Selective Forget... efficiency, content quality, and security of LLM agents
Coding patterns are bimodal: in 41% of sessions, agents author virtually all committed code ("vibe coding"), while in 23%, humans write all code themselves.
Empirical analysis of authorship attribution across the 6,000 sessions in the SWE-chat dataset; percentages derived from session-level classification.
high mixed SWE-chat: Coding Agent Interactions From Real Users in the W... distribution of code authorship across sessions (agent-dominant vs human-only se...
A determinism study of 10 replays per case at temperature zero shows both architectures inherit residual API-level nondeterminism, but DPM exposes one nondeterministic call while summarization exposes N compounding calls.
Determinism experiment with 10 replays per case at temperature zero; qualitative/quantitative observation about number of nondeterministic LLM calls exposed by each architecture.
high mixed Stateless Decision Memory for Enterprise AI Agents system nondeterminism / number of nondeterministic LLM calls exposed per decisio...
Advanced prompting methods improve accuracy on inconclusive cases but over-correct, withholding decisions even on clear cases.
Empirical comparison of prompting methods reported in paper: advanced prompts increased accuracy on inconclusive (insufficient-information) cases but led to excessive deferral/withholding on clear cases.
high mixed Learning When Not to Decide: A Framework for Overcoming Fact... accuracy on inconclusive cases and rate of withholding/deferral on clear cases
There is significant heterogeneity in methodological rigor across studies.
Authors' thematic observation from quality appraisal/extraction noting wide variation in methods, validation approaches, and reporting standards among the 64 studies.
high mixed AI-Driven Financial Risk Management and Decision Intelligenc... methodological rigor/quality of studies
AI is increasingly being integrated into both existing and newly emerging digital infrastructures, altering their architecture, functional role, and strategic significance as these systems begin to operate as embedded cognitive infrastructures shaping knowledge production, decision-making, and institutional processes.
Conceptual and descriptive claim presented by the paper (theoretical analysis/literature-informed observation). No empirical sample size or quantitative methods reported in the provided text.
high mixed Digital Sovereignty in the Global Cognitive-Informational Or... change in the architecture/role of digital infrastructures and their effect on k...
Hybrid ML+rules systems achieve partial DES-property fillability.
Result of the paper's analytic comparison across the four architectures identifying relative fillability levels for hybrid ML+rules systems.
Open-source versus closed-source trade-offs (including deployment architectures and competitive differentiation) are a central strategic consideration when selecting an enterprise LLM approach.
Paper's comparative analysis of open-source and closed-source alternatives and discussion of strategic implications; supported by the Bills Converter design rationale.
high mixed Buy Or Build? A Practitioner’s Framework for Large Language ... strategic positioning / competitive differentiation from LLM architecture choice
AI is becoming a geopolitical tool that defines trade, finance, supply chains, surveillance abilities, and diplomatic bargaining power.
Conceptual/qualitative synthesis in the paper's argument; no empirical methods or sample size reported in the abstract.
high mixed ARTIFICIAL INTELLIGENCE AND THE WEAPONIZATION OF ECONOMIC IN... influence over trade, finance, supply chains, surveillance capabilities, and dip...
The proposed safety-filter outperforms a standalone deep reinforcement learning-based controller in energy and cost metrics, with only a slight increase in comfort temperature violations.
Reported experimental comparison between the safety-filter-enhanced controller and a standalone DRL controller in the paper; specific metrics and sample size not provided in the excerpt.
high mixed Safe Deep Reinforcement Learning for Building Heating Contro... energy metrics, cost metrics, and comfort temperature violations
Confirmatory Factor Analysis (CFA) and Structural Equation Modeling (SEM) verified correlations among educational background, gender inclusiveness, digital literacy, and perceived algorithmic fairness.
Paper reports use of CFA and SEM to test relationships among those variables; reliability/fit supported by Composite Reliability (CR), Average Variance Extracted (AVE), and model-fit indicators.
high mixed A Machine Learning Perspective on FinTech-Driven Inclusion: ... correlations among educational background, gender inclusiveness, digital literac...
Benefits of technology and data analytics are context-dependent, with emerging markets facing unique regulatory and infrastructural barriers.
Narrative synthesis of included studies noting heterogeneity by context and reports of regulatory/infrastructural constraints in emerging markets.
high mixed The Use of Technology and Data Analytics in Modern Auditing:... realized benefits / adoption in varying contexts
Cybersecurity has a moderating effect on audit data analytics.
Synthesis statement in the review summarizing included studies that report cybersecurity influences the effectiveness/usability of audit data analytics.
high mixed The Use of Technology and Data Analytics in Modern Auditing:... effectiveness of audit data analytics
Digitization is reshaping the structures of Resource Dependence Theory (RDT) instead of eliminating it completely (Yordanova & Hristozov, 2025).
Conceptual/theoretical claim supported by citation to Yordanova & Hristozov (2025); presented as an interpretive conclusion about how digitization interacts with organizational dependence structures. No empirical details provided in the excerpt.
high mixed Re-Evaluation of Resource Dependence in AI Enabled SME Finan... structure of resource dependence / organizational dependence on external resourc...
Outcomes are shaped not only by benchmark quality but also by competitive pressure, including user switching, routing decisions, and operational constraints.
Argument/assertion in paper framing motivations for Marketplace Evaluation; conceptual reasoning listing mechanisms (user switching, routing, operational constraints); no empirical tests or sample size reported.
high mixed Evaluation of Agents under Simulated AI Marketplace Dynamics post-deployment system outcomes (e.g., success influenced by competition factors...
Alignment operates as a two-way translation, where models are made 'safe for worlds' while those worlds are reshaped to be 'safe for models.'
Conceptual claim supported by ethnographic examples illustrating reciprocal adaptations between models and social/institutional contexts in Nairobi's credit-scoring ecosystem.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya reciprocal adjustments between predictive models and social/institutional enviro...
Algorithmic credit scoring is accomplished through the ongoing work of alignment that stabilizes risk under conditions of persistent uncertainty, taking epistemic, modeling, and contextual forms.
The paper's theoretical argument grounded in nine-month ethnographic observations and analysis of how practitioners and institutions engage in alignment work across epistemic, modeling, and contextual dimensions.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya alignment practices that stabilize risk amid uncertainty (epistemic, modeling, c...
Practitioners negotiate model performance via technical and political means.
Observational data from the ethnography showing technical adjustments, benchmarks, and political negotiation (e.g., with regulators or management) to establish acceptable performance.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya practices used to achieve and justify model performance (technical tuning and po...
Practitioners formulate risk through multiple interpretations.
Ethnographic evidence from interviews and observations indicating that risk is characterized differently across actors (technical, legal, business interpretations).
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya variation in definitions and framings of risk among practitioners
Practitioners construct alternative data using technical and legal workarounds.
Field observations and interviews showing practitioners employing technical methods and legal strategies to create or repurpose alternative data sources for credit scoring.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya practices for generating and using alternative data in credit models
Algorithmic credit scoring is being transformed by new actors, techniques, and shifting regulations.
Ethnographic fieldwork documenting the entry of new actors, novel technical techniques, and regulatory changes affecting credit scoring in Nairobi's digital lending ecosystem.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya structural transformation of algorithmic credit scoring (actor composition, tech...
Credit scoring is an increasingly central and contested domain of data and AI governance.
Nine-month ethnography of credit scoring practices in Nairobi, Kenya; participant observation and interviews across stakeholders in digital lending.
high mixed Risk, Data, Alignment: Making Credit Scoring Work in Kenya role of credit scoring in data and AI governance (centrality and contestedness)
The local labor market will follow a dual trajectory: low-skill, routine jobs face high automation risk while demand will rise for AI-collaborative, higher-skill roles.
Paper's analytical prediction based on distinguishing current job roles into routine/repetitive vs cognitive/non-routine and projecting likely impacts; no numeric forecasts or sample sizes provided in the excerpt.
high mixed PREDICTING THE FUTURE OF JOBS IN NAGPUR DISTRICT MIDC: THE R... combined job displacement for routine roles and increased demand for AI-collabor...
Professional and Technical Services, Information, and Finance and Insurance account for approximately 86 percent of the base-case direct contribution.
Sectoral decomposition of base-case direct contribution in the model; paper explicitly reports the three sectors' combined share as ~86%.
high mixed AI Capex Is Justified: A Bottom-Up Sectoral Estimate of Arti... share of base-case direct GDP contribution by sector (three-sector concentration...
The inverted U-shaped pattern between AI knowledge stickiness and technological concentration is more clearly detected in eastern cities and in small and medium-sized cities; in large cities the quadratic term is not statistically significant.
Heterogeneity/subsample regressions by region (east vs. other) and city size categories within the city-year panel (2014–2023); statistical significance of quadratic term differs across subsamples.
high mixed Knowledge stickiness and technological concentration in the ... technological concentration (presence and significance of nonlinear relationship...
Technological complexity moderates the nonlinear (inverted U) association between AI knowledge stickiness and technological concentration by altering its strength and curvature rather than producing a simple, uniform shift in the turning point.
Interaction/heterogeneity analyses in the two-way fixed-effects city-year panel (2014–2023), examining moderating role of a technological complexity measure on the quadratic association.
high mixed Knowledge stickiness and technological concentration in the ... technological concentration (degree and curvature of the stickiness–concentratio...
There is an inverted U-shaped association between AI knowledge stickiness and technological concentration: higher stickiness up to a limit leads to more concentration and thereafter the opposite.
City-year panel combining AI patent applications with urban statistics for 2014–2023; two-way fixed-effects regression showing a significant positive linear and negative quadratic term (nonlinear association).
high mixed Knowledge stickiness and technological concentration in the ... technological concentration (allocation of AI activity across sub-technology bra...
Subjectivity persisted in AI-powered recruitment decisions; human judgment remained an important factor.
Theme 2 (subjectivity in AI-powered recruitment) from interviews indicating retained human subjectivity and judgement in recruitment processes (n = 22).
high mixed The augmented recruiter: examining AI integration and decisi... degree_of_subjectivity_in_decision_making
Big data analytics (BDA) adoption is a risky strategy with potentially high rewards for start-ups.
Stated as a summary conclusion based on empirical analysis of a large sample of start-ups in Germany comparing adopters and non-adopters across multiple performance measures (survival, costs, sales, employee growth, access to financing).
high mixed Big data-based management decisions and start-up performance overall performance/risk–reward tradeoff
Bounded agents act as an amplifying but not necessary extension to the foundation-model stack for changing work coordination.
Conceptual argument within the paper distinguishing bounded agents from the core stack; no empirical comparison or measurement reported.
high mixed Remote-Capable Knowledge Work Should Default to AI-Enabled F... role of bounded agents in amplifying coordination impacts
The spatial spillover effects are geographically constrained and vary significantly across regions.
Reported heterogeneity in spatial Durbin model results and discussion of geographic constraint and inter-regional variation (regional heterogeneity analysis).
high mixed Research on the Pathways and Spatial Effects of Digital–Inte... heterogeneity of spatial spillover effects on carbon intensity across regions
The effects of generative AI on work and organisations are heterogeneous and context-dependent, shaped by job roles, skill levels, and institutional environments.
Synthesis across the included studies noting variation in outcomes conditional on role, skill, and institutional context.
high mixed Generative AI in the Workplace: A Systematic Review of Produ... heterogeneity of AI effects across roles/skills/institutions
Overall, AI emerges as a transformative but context-dependent tool for business decision-making in Latin America.
The authors' overall interpretation and synthesis of the 27 reviewed studies highlighting variable outcomes depending on context and readiness.
high mixed Artificial Intelligence for Business Decision-Making in Lati... overall impact of AI on business decision-making (transformative effect conditio...
Although the concurrent paradigm performs worse than the sequential paradigm in terms of immediate task performance, it is more effective in promoting users' emotional trust.
Comparison between concurrent and sequential AI-assisted decision-making paradigms in the RCT (N=120); authors report concurrent < sequential for immediate task performance, but concurrent > sequential for emotional trust.
high mixed How AI-Assisted Decision-Making Paradigms and Explainability... immediate task performance (negative) and emotional trust (positive)
AI adoption outcomes depend on organizational routines, data arrangements, accountability structures, and public values.
Empirical and theoretical literature review and argument in the article drawing on scholarship in digital government and public-sector technology adoption.
high mixed Governing frontier general-purpose AI in the public sector: ... determinants of AI adoption in government (organizational, data, accountability,...
If employment losses are relatively small and productivity gains are realised, AI adoption could boost Exchequer revenues. But if job displacement is sizeable, tax receipts fall while welfare spending rises, resulting in potentially large pressures on the public finances.
Conditional fiscal scenarios simulated in the report combining employment, wage and benefit changes with the public finance implications (tax receipts and welfare spending); reported as scenario-based outcomes.
high mixed Artificial Intelligence and income inequality in Ireland Exchequer revenues / tax receipts and welfare spending
Ireland’s tax and welfare system absorbs most of the income loss for lower income households, and roughly half of the loss for households at the top of the income distribution.
Microsimulation using SWITCH to model taxes and transfers applied to simulated income changes across income groups; reported as a finding in the report.
high mixed Artificial Intelligence and income inequality in Ireland net income after taxes and transfers (absorption of income loss)
India exhibits a distinctive polarisation pattern: a shrinking middle-skill workforce alongside a persistently large low-skill labour segment.
Descriptive analysis of secondary data and official reports from 2020–2024 comparing occupational and skill distributions in India.
high mixed Artificial Intelligence and labour market polarisation in In... changes in the share of labour across skill bands (middle vs low skill)
Mathematics (SAFI: 73.2) and Programming (71.8) receive the highest automation feasibility scores; Active Listening (42.2) and Reading Comprehension (45.5) receive the lowest.
SAFI benchmark results reported for specific O*NET skills (numerical SAFI scores provided in the paper).
high mixed The AI Skills Shift: Mapping Skill Obsolescence, Emergence, ... SAFI score by skill (automation feasibility)
Only a small subset of LLM retailers can consistently achieve capital appreciation, while many hover around the break-even point.
Empirical results from the 20-agent benchmark experiments reported in the paper, contrasting capital appreciation for winners vs break-even for many agents.
high mixed Market-Bench: Benchmarking Large Language Models on Economic... capital appreciation / agent profitability
Benchmarking on 20 open- and closed-source LLM agents reveals significant performance disparities and a winner-take-most phenomenon.
Empirical evaluation described in the paper using 20 LLM agents (open- and closed-source); results reported show uneven performance distribution.
high mixed Market-Bench: Benchmarking Large Language Models on Economic... performance (financial/competitive outcomes of retailer agents)
Tool developers, users, and social scientists conceptualize 'context' differently, and these divergent conceptualizations reveal specific pitfalls inherent in computational approaches to context.
Analytic comparison across stakeholder perspectives derived from interviews and conceptual analysis in the paper (qualitative evidence; sample size unspecified).
high mixed Context Collapse: Barriers to Adoption for Generative AI in ... differences in conceptual definitions and the resulting pitfalls for computation...
AI adoption significantly reshaped task profiles for 73% of respondents, particularly affecting routine data processing, administrative tasks, and scheduling activities.
Survey data and secondary data analysis reported in this study (sample size not stated); self-reported change in task profiles with reported percentage (73%).
high mixed Artificial Intelligence Adoption and Career Reconfiguration ... task profile change (impact on routine data processing, administrative tasks, sc...
There is a robust inverted U-shaped relationship between robotics manufacturing development and urban carbon emissions.
Panel data analysis using 277 Chinese prefecture-level cities from 2008 to 2019; econometric analysis reported in the paper finds an inverted U-shaped association and robustness checks are claimed.
AI adoption across firms is heterogeneous, varying across sectors such as finance, technology, and manufacturing.
Survey of 150 leading Nigerian firms across finance, tech, and manufacturing showing variation in AI integration; supported by qualitative interviews and policy analysis.
high mixed Human Capital and the AI-Powered Future of Work: (Training, ... heterogeneity in AI adoption across firms/sectors
The rapid, heterogeneous integration of Artificial Intelligence (AI) technologies is profoundly reshaping the dynamics of work across the Nigerian business sector, generating both significant economic opportunities and acute labor market challenges.
Mixed-methods study combining a quantitative survey of 150 leading Nigerian firms across finance, tech, and manufacturing and qualitative analysis of government policy and workforce interviews.
high mixed Human Capital and the AI-Powered Future of Work: (Training, ... dynamics of work (economic opportunities and labor market challenges)
Both rapid model improvement and benchmark quality issues contributed to underestimating agent capabilities.
Synthesis of results: improved LLM performance plus audit findings showing benchmark errors together explain the prior underestimation; based on the re-evaluation and audit described in the paper.
high mixed ELT-Bench-Verified: Benchmark Quality Issues Underestimate A... factors contributing to underestimation of agent capabilities (model improvement...
Models performed well on commonly discussed topics but struggled with specialized health data.
Task-level performance comparison across topics in the elicited population statistics: better accuracy on commonly discussed topics, poorer performance on specialized health data tasks.
high mixed Bayesian Elicitation with LLMs: Model Size Helps, Extra "Rea... topic-specific estimation accuracy
In a preliminary experiment, giving models web search access degraded predictions for already-accurate models, while modestly improving predictions for weaker ones.
A preliminary comparative test where some models were given web search access and changes in predictive performance were observed: degradation for already-accurate models and modest improvement for weaker models.
high mixed Bayesian Elicitation with LLMs: Model Size Helps, Extra "Rea... change in predictive accuracy with web search access