Evidence (5267 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 |
Adoption
Remove filter
Because the sample is non-representative (support-group recruitment and media cases) and small (19 users), the authors note that generalizability is limited and the sample is biased toward more severe cases.
Limitations section stating recruitment sources, small N, and bias toward severe cases.
The study analyzed conversation logs from 19 users who reported psychological harm associated with chatbot use, comprising a total corpus of 391,562 messages (user + chatbot).
Dataset described in paper: 19 users' conversation logs aggregated; total message count reported as 391,562 messages across user and chatbot messages.
Metrics used to evaluate agents include operational stability (e.g., variance or frequency of catastrophic failures), efficiency (e.g., cost/profit/fulfillment), and degradation across increasing task complexity.
Methods and experimental sections specifying the metrics applied to compare ESE and baselines on RetailBench environments.
Baselines used in comparisons include monolithic LLM agents and other existing agent architectures that do not implement explicit strategy/execution separation.
Experimental design: baseline descriptions in the methods section specifying monolithic LLM agents and additional architectures lacking explicit temporal decomposition.
Eight state-of-the-art LLMs were evaluated in the study.
Experimental setup description listing eight contemporary LLMs tested across RetailBench environments.
The paper proposes Evolving Strategy & Execution (ESE), a two-tier architecture that separates high-level strategic reasoning (updated at a slower temporal scale) from low-level execution (short-term action selection).
Architectural design described in the methods: explicit decomposition into strategy and execution modules with differing update cadences and stated interpretability/adaptation mechanisms.
RetailBench environments are progressively challenging to stress-test adaptation and planning capabilities (i.e., environments increase in complexity, stochasticity, and non-stationarity).
Benchmark construction described in the paper: multiple environment difficulty levels used to evaluate degradation under increasing challenge; experiments run across these progressive environments.
The paper introduces RetailBench, a high-fidelity long-horizon benchmark for realistic commercial decision-making under stochastic demand and evolving external conditions (non-stationarity).
Design and presentation of the benchmark in the paper: simulated commercial operations with stochastic demand processes and shifting external factors; emphasis on long-horizon evaluation and progressively challenging environments.
Roughly 25% of the training corpus is Italian-language data.
Corpus composition reported by the authors: Italian-language share ≈25% of total training tokens. The summary cites this proportion but does not list the datasets or language-detection methodology.
The model was trained on approximately 2.5 trillion tokens of data.
Training-data size reported in the paper (aggregate token count ≈2.5T). The summary provides this number; no per-dataset breakdown or provenance details are included in the summary.
Approximately 3 billion parameters are active per inference (sparse activation / ~3B active parameters at runtime).
Paper reports sparse MoE design with ≈3B active parameters per forward pass. Evidence comes from model design description (active set / routing), not from independent runtime FLOP logs in the summary.
EngGPT2-16B-A3B is a Mixture-of-Experts (MoE) model trained from scratch with a total of 16 billion parameters.
Model specification reported in the paper: architecture described as MoE and total parameter count listed as 16B. No contrary empirical test needed; claim is a declarative model spec.
The project developed domain- and specialty-focused models: Fanar-Sadiq (Islamic content multi-agent architecture), Fanar-Diwan (classical Arabic poetry), and FanarShaheen (bilingual translation).
Paper enumerates these domain/specialty models and their stated focuses as part of the product stack.
FanarGuard is a 4B bilingual moderation model focused on Arabic safety and cultural alignment.
Paper lists FanarGuard in the expanded product stack and specifies model size (4B) and bilingual moderation purpose emphasizing Arabic safety/cultural alignment.
Fanar-27B was produced by continual pre-training from a Gemma-3-27B 27B backbone.
Paper describes model development: continual pre-training of Fanar-27B from the Gemma-3-27B 27B backbone.
The Fanar 2.0 training corpus is a curated set totalling approximately 120 billion high-quality tokens organized into three data 'recipes' emphasizing Arabic and cross-lingual relevance.
Paper reports a curated corpus of ~120B high-quality tokens split across three data recipes; emphasis on relevance and quality for Arabic and cross-lingual performance.
Training and operations for Fanar 2.0 were performed on-premises using 256 NVIDIA H100 GPUs at QCRI.
Paper states compute and infrastructure: training and operations performed on 256 NVIDIA H100 GPUs, fully on-premises at QCRI (HBKU).
Experiments were conducted on three benchmarks and across multiple LLM families to evaluate generation, scoring, calibration, robustness, and efficiency dimensions.
Data & Methods section summary in the paper stating systematic evaluation across three benchmarks and a variety of LLMs and verifiers.
Complete provenance of training data is often unavailable, so contamination detection is imperfect and some leakage may be undetectable (or overestimated in some categories).
Authors' stated limitation about unavailable/partial training-data provenance and methodological caveats for the lexical-matching pipeline and behavioral probes.
Results are specific to MMLU; contamination levels and effects may differ on other benchmarks or newer models.
Authors' limitations: experiments were conducted only on the MMLU dataset (513 questions) and on the listed six models; generalizability is therefore uncertain.
A three-layer evaluation framework was applied systematically: Layer 1 = syntactic validity; Layer 2 = semantic correctness; Layer 3 = hardware executability (with sublayer 3b = end-to-end evaluation on quantum hardware).
Methods section describes application of a three-layer evaluation framework to each reviewed system, including the explicit sublayer 3b definition.
The review grouped training regimes across the systems as supervised fine-tuning, verifier-in-the-loop reinforcement learning (RL), diffusion/graph generation, and agentic optimization.
Surveyed systems' training descriptions were classified into these training-regime categories during the review's analytical synthesis.
The review organized artifacts along artifact-type axes: Qiskit code, OpenQASM programs, and circuit graphs.
Analytical organization described in the methods: artifact-type axis enumerated as Qiskit, OpenQASM, and circuit graphs across the surveyed systems.
"Quantum code" in this review is defined as program artifacts (Qiskit code, OpenQASM); quantum error-correcting code (QEC) generation was excluded.
Inclusion/exclusion criteria specified in the review explicitly limited scope to program artifacts such as Qiskit and OpenQASM and excluded QEC-focused works.
A structured scoping review (Hugging Face, arXiv, provenance tracing; Jan–Feb 2026) identified 13 generative systems and 5 supporting datasets relevant to quantum circuit / quantum code generation.
Structured search of Hugging Face model/dataset listings, arXiv literature, and provenance tracing conducted between January and February 2026; results yielded 13 systems and 5 datasets (sample counts reported in the review).
The reinforcement learning objective optimizes a combined utility that trades off task success and resource costs; the reward penalizes delays and failures.
Learning method section describes training the high-level orchestrator with an RL reward that penalizes delays (latency/resource consumption) and failures, and that algorithmic/hyperparameter details are provided.
The experiments use empirical LLM latency profiles measured from ALFRED tasks to model realistic inference delays in simulation.
Environment/evaluation description states use of an embodied task suite based on ALFRED and empirical latency profiles to model realistic LLM inference delays.
Baselines for comparison include fixed reasoning strategies (always reason, never reason), heuristic triggers for invoking LLMs, and ablations of RARRL components.
Paper lists these baselines explicitly in the Baselines and comparisons section and reports experiments comparing RARRL to them.
The high-level orchestration policy uses observations that include current sensory observation, execution history, and remaining resources (e.g., remaining time or compute budget).
Key Points and Methods specify the observation space used by the orchestrator, listing sensory inputs, execution history, and resource remaining as inputs.
RARRL trains only a high-level orchestration policy via reinforcement learning and does not retrain the existing low-level control/policy modules end-to-end.
Methods/Model architecture describe a hierarchical approach where low-level controllers are existing modules and are not retrained; RL is applied to the high-level orchestrator.
RARRL (Resource-Aware Reasoning via Reinforcement Learning) is a hierarchical orchestration framework that learns a high-level policy to decide when an embodied agent should invoke LLM-based reasoning, which reasoning role to use, and how much compute budget to allocate.
Paper describes a hierarchical design with a learned high-level RL orchestrator that issues discrete decisions about reasoning invocation, reasoning role/mode, and compute budget allocation; architecture and decision space specified in Methods.
BenchPreS defines two complementary metrics—Misapplication Rate (MR) and Appropriate Application Rate (AAR)—to quantify over‑application and correct personalization, respectively.
Methodological contribution described in the paper: explicit definitions of MR as fraction of inappropriate applications and AAR as fraction of appropriate applications, used to score model behavior.
Pilot randomized or quasi-experimental implementations of reduced workweeks (across firms, industries, or regions) are needed to measure effects on employment, productivity, wages, and consumption.
Research-design recommendation motivated by lack of contemporary causal evidence; not an empirical finding but a stated priority for rigorous testing.
There is limited direct causal identification separating technology-driven layoffs from incentive-driven layoffs in current firm-level data, creating a need for new firm-panel datasets linking AI adoption, executive pay/ownership, layoff decisions, and local demand outcomes.
Stated limitation of the paper and research-priority recommendation; assessment based on literature gaps noted in the synthesis rather than empirical gap quantification.
Observed layoffs should be treated in empirical research as outcomes of firm governance and incentive structures; econometric studies estimating displacement from AI must control for managerial incentives and financial pressures.
Methodological recommendation based on the conceptual argument and literature linking governance/incentives to firm behavior; no new empirical demonstration provided.
Key empirical metrics introduced and used are: AI adoption rates (sector-level intensity), Skill shift index, Hybrid job share, and employment levels/net changes by sector.
Methods description listing the constructed metrics used in the simulated dataset and subsequent analyses (definitions and calculation procedures provided in the paper).
The study's main limitations include reliance on a simulated dataset rather than exhaustive administrative microdata, literature limited to selected publishers/years, and correlational (not causal) identification of some effects.
Authors' explicitly stated limitations in the paper's methods and discussion sections describing data choices (simulated dataset, selected publishers 2020–2024) and the observational/correlational nature of several analyses.
Further research is needed—randomized controlled trials, long-term impact measurement (earnings, employment stability, skill accumulation), distributional analysis, and model audits for bias.
Authors' stated research agenda and recommendations; not an empirical finding but a methodological recommendation following the pilot.
The authors explicitly note limitations: the study focuses on prediction (not causation), results are sensitive to data quality, workforce records may contain biases, and practical constraints like privacy and deployment complexity limit direct operational adoption.
Limitations section described by the authors listing prediction-versus-causation distinction, sensitivity to data quality, potential biases, privacy concerns, and deployment complexity.
The study used a reproducible modeling pipeline (data cleaning, feature engineering, model training and tuning, systematic evaluation) applied to several freely available workforce datasets to enable replication.
Methods section describes a reproducible workflow including preprocessing steps, engineered features, hyperparameter tuning for each model class, cross-validation, and use of publicly available datasets.
This work is conceptual/theoretical and reports no original empirical dataset; it explicitly calls for mixed-methods empirical validation (case studies, field experiments, longitudinal studies), measurement development, and multi-level data collection.
Explicit methodological statement in the paper describing its nature as a theoretical synthesis and listing empirical needs; no empirical sample provided.
The paper recommends an empirical research agenda including field experiments comparing teams with and without AI mediation, structural models of labor supply and wages under reduced language frictions, microdata analysis of adopters, and measurement studies for coordination costs and mediated-action reliability.
Explicit recommendations and research agenda stated in the paper; this is a descriptive claim about the paper's content rather than an empirical finding.
The paper's primary approach is conceptual/theoretical development and agenda-setting; it does not report large-scale empirical or experimental data.
Explicit methods statement in the paper: synthesis, illustrative examples, framework development; absence of reported empirical sample or experiments.
The study's empirical base consists of 40 semi-structured interviews with cross-industry project practitioners in the UK, analyzed using thematic qualitative methods.
Stated data and methods in the paper: sample size (40), interview method, cross-industry sampling, and thematic analysis.
Limitation: Implementation heterogeneity — the costs and feasibility of the recommended HR changes vary by context and may affect generalisability.
Explicit limitation acknowledged in the paper; drawn from theoretical reasoning about contextual heterogeneity and practitioner variability.
Limitation: The framework is conceptual and requires empirical validation across sectors, firm sizes and AI‑intensity levels.
Explicit limitation acknowledged by the authors; based on the paper's method (theoretical synthesis, no original data).
The paper generates empirically testable propositions (e.g., how leader practices affect AI adoption speed, task reallocation, productivity, error rates, employee well‑being and turnover) and suggests natural‑experiment settings for evaluation.
Stated methodological output of the conceptual synthesis; the paper lists candidate empirical tests and research opportunities but contains no original empirical tests.
Typical methods used are deep learning for property prediction and representation learning, protein-structure modelling tools, generative models for de novo design, NLP for knowledge extraction, and ADME/Tox in silico models integrated with traditional computational chemistry.
Methodological survey in the paper listing these approaches and examples of their application.
Commonly used data types in AI-driven drug discovery include biochemical/binding assay data, protein structural data, HTS results, ADME/Tox and PK datasets, omics/phenotypic readouts, and scientific literature/patents.
Cataloguing of data sources used across studies and company pipelines described in the paper.
AI became widely adopted in pharmaceutical discovery during the 2010s, driven by greater compute, larger datasets, and advances in deep learning.
Historical overview and trend analysis in the paper referencing increased compute availability, growth in public and proprietary datasets, and the rise of deep-learning publications and tools over the 2010s.