Evidence (3308 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
Filter claims →
Org Design
4892 claims
Filter claims →
Innovation
4781 claims
Filter claims →
Labor Markets
4004 claims
Filter claims →
Skills & Training
3308 claims
Filtered →
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 |
Skills Training
Remove filter
To preserve long-term resilience, engineering leaders must move beyond prompt-based development to implement rigorous human-in-the-loop pedagogical standards.
Prescriptive recommendation based on the paper's conceptual analysis; no randomized trials or empirical validation of this intervention reported in the excerpt.
Grounding recommendations in validated research offers leaders a framework for navigating AI's labor implications responsibly.
Paper asserts that its synthesis and recommendations provide a practical framework for leaders; no empirical validation of the framework is reported in the abstract.
Evidence-based organizational responses (transparent workforce planning, skills investment, redesigned roles, adaptive governance, and long-term capability-building) can mitigate harm and prepare organizations for workplace transformation.
Paper proposes these organizational responses grounded in the synthesized empirical literature; this is a recommendation rather than an empirically tested intervention in the paper abstract.
There is an absence of a comprehensive national strategy in Israel for AI in employment, and the paper calls for the development of a forward-looking regulatory framework that balances innovation with protection of fundamental rights (dignity, equality, privacy), transparency, human oversight, and fairness.
Normative policy recommendation based on the paper's regulatory analysis; not an empirical finding and no policy-design experiments are reported in the excerpt.
The AI-driven transformation is accompanied by an increasing emphasis on reskilling and continuous learning, reflecting a shift from workforce replacement to reconfiguration of modes of employment.
Reported observation in the paper about workforce development trends; no quantitative measures of reskilling uptake or program counts are provided in the excerpt.
Israeli legal scholarship reflects broad interdisciplinary engagement with AI across labor law, intellectual property, privacy, constitutional law, and additional fields; the study advances theoretical models, including reconceptualizations of accountability, creativity, and the role of AI as a legal actor.
Literature review/academic survey and theoretical contributions reported in the paper; specific counts of publications or analytical methods not provided in the excerpt.
Israel is a leading “AI Nation,” characterized by exceptionally high levels of technological integration across both the private and public sectors.
Statement in paper based on the author's characterisation of national-level technological integration; specific empirical measures or sample size not provided in the excerpt.
For software engineers, GAI's (GHC's) productivity impacts and creation of new tasks appear to outweigh potential displacement effects from automation of some SWE tasks.
Interpretation based on observed associations: higher hiring probability (especially entry-level), increased non-programming skills in new hires, and no decline in coding skills in the LinkedIn/GitHub observational data.
New hires at GHC-adopting firms exhibit around 5% more non-programming skills.
Analysis of LinkedIn skill listings for new hires linked to GitHub/GHC adoption status, comparing the prevalence/count of non-programming skills among new hires at adopters versus non-adopters.
The increase in hiring probability is driven by entry-level hires.
Subgroup/heterogeneity analysis within the LinkedIn/GitHub observational data showing the hiring increase concentrated among entry-level SWE hires.
GHC adoption is associated with around a 3%–5% higher monthly probability of hiring SWEs.
Observational analysis using LinkedIn and GitHub data comparing firms that adopted GitHub Copilot (GHC) to firms that did not; association measured as change in firms' monthly probability of hiring software engineers.
Technological advancement alone is insufficient—maximizing AI's economic potential requires strategic investments in workforce capability development (e.g., widespread AI fluency programs and targeted cultivation of higher-order judgment skills).
Policy recommendation based on the article's synthesis of task-based models and empirical literature; the excerpt does not report specific interventions, trials, or sample sizes.
The supply of AI-literate workers amplifies productivity gains.
Stated as a mechanism in the task-based model synthesis; described qualitatively in the article without specific empirical method or sample sizes in the excerpt.
Aggregate productivity improvements from AI advancement depend critically on two forms of human capital: specialized AI expertise and complementary non-AI skills.
Claim is presented as a theoretical result drawn from 'task-based economic models' in the article; empirical corroboration is referenced generally but no specific datasets or sample sizes are reported in the excerpt.
Mounting empirical evidence indicates AI primarily functions as augmentation technology—amplifying human capabilities rather than replacing workers.
Article states it draws on 'mounting empirical evidence' and synthesizes recent theoretical and empirical findings; no specific studies, methods, or sample sizes are cited in the excerpt.
Im Forschungskontext sind kontextbezogene Schulungs- und Begleitmaßnahmen entscheidend für den Erfolg der Copilot-Einführung.
Schlussfolgerung der Autoren aus den Befunden zur zeitlichen Entwicklung der Bewertungen wissenschaftlicher Mitarbeitender und zu unterschiedlichen Nutzenwahrnehmungen (im Abstract genannt).
Die Untersuchung zeigt, dass Microsoft 365 Copilot insbesondere im administrativen Bereich Effizienzgewinne ermöglicht.
Selbstberichtete Einschätzungen der Beschäftigten (speziell Verwaltungsmitarbeitende) in der wiederholten Querschnittsbefragung; Autoren ziehen daraus praktische Relevanz im administrativen Bereich (Abstract).
Die Befunde unterstreichen die Bedeutung kontextspezifischer Einführung, rollenbezogener Qualifizierung und Governance für eine nachhaltige Akzeptanz generativer KI in Organisationen.
Interpretation/Schlussfolgerung der Autoren basierend auf den survey-Ergebnissen und beobachteten Unterschieden zwischen Rollen sowie zeitlichen Entwicklungen (im Abstract formuliert).
Der größte Mehrwert von Copilot liegt bei klar strukturierten, textbasierten Aufgaben.
Befragungsergebnisse zur Nutzenabschätzung für typische Tätigkeiten der Wissensarbeit, wie im Abstract zusammengefasst (präferierte Aufgabenarten: strukturierte, textbasierte Aufgaben).
Microsoft 365 Copilot wird überwiegend als benutzerfreundlich und technisch zuverlässig wahrgenommen.
Selbstberichtete Beurteilungen zu Benutzerfreundlichkeit und technischer Zuverlässigkeit in der wiederholten Querschnittsbefragung (Angabe im Abstract).
Wissenschaftliche Mitarbeitende entwickeln im Zeitverlauf positivere Einschätzungen, insbesondere hinsichtlich Produktivität und Arbeitserleichterung durch Copilot.
Längsschnittähnliche Beobachtung über die wiederholten Querschnittserhebungen; zeitliche Veränderung der Selbsteinschätzungen wissenschaftlicher Mitarbeitender im Abstract beschrieben.
Verwaltungsmitarbeitende bewerten die Nützlichkeit und die Zuverlässigkeit von Microsoft 365 Copilot höher als wissenschaftliche Mitarbeitende.
Selbstberichtete Bewertungen in der wiederholten Querschnittsbefragung; Vergleich zwischen Berufsrollen (Verwaltung vs. Wissenschaft) angegeben im Abstract.
Educators, policymakers, and industry leaders should design AI-inclusive curricula, workforce development strategies, and policies that support sustainable human–AI collaboration.
Policy and practice recommendations derived from the review's synthesis of empirical findings and identified gaps; presented as conclusions and directions.
AI is not simply replacing jobs but is redefining professional identity in IT, emphasizing reskilling, adaptability, and lifelong learning as key determinants of future employability.
Synthesis of reviewed literature and the paper's concluding interpretation summarizing trends across empirical studies, industry reports and conference findings.
There is growing demand for hybrid skill sets that integrate technical expertise with higher-order cognitive, ethical, and socio-emotional competencies among IT professionals.
Reported across reviewed empirical studies and industry reports summarized in the review paper.
When firms adopt AI as an augmentative tool rather than a replacement mechanism, it can raise worker productivity and contribute to job creation.
Literature review citing empirical examples and studies of AI augmentation that increased productivity and produced new job roles (empirical studies summarized).
Combining insights from multiple disciplines, the review contributes to broader discussions on creating AI-enabled work environments that are both innovative and gender-inclusive.
Stated as the paper's contribution and framing in the abstract; based on the paper's described interdisciplinary literature synthesis rather than new empirical findings.
Practical recommendations that improve gender-inclusive outcomes include reskilling, mentorship programs, bias-aware AI deployment, and inclusive organizational design.
Recommendations synthesized from the reviewed literature and policy analyses; the abstract does not indicate rigorous causal evaluations or quantification of the effectiveness of these interventions within the paper.
There exist successful initiatives, organizational strategies, and policy interventions that have enhanced women’s inclusion, career progression, and representation in emerging tech roles.
Paper reports examples from the reviewed literature and policy analyses that are characterized as 'successful initiatives'; the abstract does not list specific programs, evaluation designs, or sample sizes.
This work contributes by integrating fragmented literature into a coherent, comparative perspective that offers actionable insights for researchers, policy makers, and industry stakeholders.
Author claim about the contribution of the review (self-assessment; no external validation reported in the abstract).
Findings highlight the growing importance of re-skilling and adaptive policy measures to mitigate employment risks associated with AI.
Policy recommendation derived from the review and synthesis of sectoral literature (no empirical evaluation of re-skilling program effectiveness provided in the abstract).
Knowledge-driven domains experience significant augmentation and skill shifts rather than displacement.
Reported synthesis from the systematic review comparing sectoral effects (qualitative statement; no quantified effect sizes or counts in the abstract).
Results indicate that AI-assisted text-to-model methods can substantially lower the cost of constructing structured procedural representations, making course-wide deployment of structured AI coaching systems practically feasible.
Conclusion drawn from reported results (e.g., time reductions and modeled outputs); the paper claims that these results imply lower costs and practical feasibility for course-wide deployment.
AI-assisted authoring reduced expert modeling time by 50–70% while producing structurally valid and highly reproducible models under fixed-input conditions.
Quantitative claim reported in the paper comparing expert modeling time with AI assistance and reporting structural validity and reproducibility under fixed-input conditions; exact experimental setup and sample size not stated in the abstract.
We apply the pipeline to instructional materials from a graduate-level online AI course, constructing 23 procedural skill models.
Empirical application reported in the paper: the pipeline was run on course materials and produced 23 models (number explicitly stated).
The approach automates structural scaffolding while preserving expert oversight for validating causal transitions and failure conditions.
Claim about system design and human-in-the-loop workflow reported in the paper; implies human validation steps are maintained alongside automated generation.
We present a human-in-the-loop text-to-model pipeline that uses large language models to transform instructional materials into schema-complete Task-Method-Knowledge models via ontology-constrained prompting and template-based generation.
Methodological contribution described in the paper; pipeline design and implementation reported (no separate quantitative validation in this sentence).
Humble leadership indirectly alleviates the negative indirect effect of HAI-C task complexity on work engagement by enhancing employees' AI self-efficacy.
Reported moderated mediation/conditional process findings from hierarchical regression and bootstrapping on the three-wave matched sample of 497 employees.
AI self-efficacy mitigates (buffers) the negative indirect impact of HAI-C task complexity on employees' work engagement.
Moderated mediation analysis conducted on longitudinal survey data (n=497) using hierarchical regression and bootstrapping; reported in Results that AI self-efficacy weakens the negative indirect effect.
HAI-C task complexity increases employees' HAI-C tech-learning anxiety.
Longitudinal survey data (n=497) analyzed with hierarchical regression; reported as a finding in the Results that task complexity amplifies tech-learning anxiety.
SWE-chat is a living dataset; our collection pipeline automatically and continually discovers and processes sessions from public repositories.
Description of the dataset collection infrastructure and pipeline provided in the paper; operational behavior asserted by authors.
The dataset currently contains 6,000 sessions, comprising more than 63,000 user prompts and 355,000 agent tool calls.
Descriptive statistics reported by the authors based on their dataset collection pipeline (dataset metadata).
We present SWE-chat, the first large-scale dataset of real coding agent sessions collected from open-source developers in the wild.
Paper authorship / dataset description; dataset curated and presented by the paper as a contribution. No external validation provided in excerpt.
We propose seven interface primitives operationalizing verification-centered HCI.
Design contribution: specification of seven interface primitives within the paper (conceptual/design proposal); no user-study or empirical validation reported.
We map synthetic literacy -- oral input generating literate output -- as the defining feature of this transition.
Conceptual mapping and theoretical framing within the paper; supported by examples from technology trends but no empirical evaluation reported.
Knowledge workers become adversarial auditors rather than keystroke-producers.
Projected role-shift based on the verification-bottleneck thesis and interdisciplinary supporting arguments; no empirical longitudinal workforce study reported.
The central contribution identifies the verification bottleneck: as AI collapses production friction, the primary constraint shifts from generation to evaluation.
Theoretical argument supported by literature synthesis across multiple fields; no direct experimental quantification provided.
Code-generating Artificial Intelligence has gained popularity within both professional and educational programming settings over the past several years.
Background statement in the paper's introduction (observational claim about recent trends in AI adoption).
The emotional effect of the human teammate was significantly more positive and arousing compared to working with Copilot.
Subjective emotion measures (valence/arousal) collected in the study; reported significant differences favoring human teammate on positivity and arousal (n=22).
Several dimensions of participants' workload were significantly reduced when using GitHub Copilot.
Subjective workload measures collected during the experiment; multiple workload dimensions reported as significantly lower in the Copilot condition (n=22).