Aug 2026· Magna Scientia Advanced Research and Reviews· Vol 17, pp. 446-457· 0 citations
TL;DR
This paper articulates an integrated nationwide framework featuring federated data interoperability, risk-calibrated algorithmic hygiene standards, mandatory prospective evaluation protocols, and a phased evolution toward binding administrative regulation.
Abstract
The integration of predictive analytics into United States occupational safety and health practice promises to shift workplace risk management from post-incident recordkeeping toward prospective injury prevention. Synthesizing U.S.-focused research published between 2021 and 2026, this narrative review critically evaluates machine learning applications across severe incident classification, narrative text processing, real-time computer vision, return-to-work outcome forecasting, and emerging federal oversight models. While algorithmic capabilities have reached high computational performance, including production-scale transformer deployments for administrative coding and high-accuracy ensembles for accident narrative parsing, the literature remains dominated by retrospective offline experiments. Crucially, empirical evaluations demonstrate a profound gap between model precision and tangible worker safety, as virtually no published studies measure prospective reductions in workplace injury or illness rates. This lack of demonstrated field impact is further complicated by severe systemic data fragmentation across federal enforcement registries, statistical surveys, sector-specific databases, and state-bounded workers' compensation claims. To bridge this divide, this paper articulates an integrated nationwide framework featuring federated data interoperability, risk-calibrated algorithmic hygiene standards, mandatory prospective evaluation protocols, and a phased evolution toward binding administrative regulation. Aligning computational innovation with measurable workplace hazard reduction, rather than further optimizing classification accuracy on historical datasets, represents the essential mandate for the future of occupational safety analytics.
Occupational injury and disease surveillance plays an important role in supporting prevention-oriented occupational safety and health (OSH) governance. However, under-reporting of occupational injuries and diseases remains a persistent challenge, particularly in emerging economies where surveillance systems are often fragmented and administratively oriented. This study examines potential under-reporting patterns within Indonesia's occupational surveillance system across five industrial sectors, manufacturing, mining, palm oil, fisheries, and telematics, during 2019–2024. The study employed a quantitative analytical design integrating administrative claims analysis, Random Forest classification, K-Means clustering, and regulatory review using de-identified data from BPJS Ketenagakerjaan. The findings reveal substantial disparities between occupational injury (JKK) and occupational disease (PAK) reporting across sectors. Manufacturing recorded the highest occupational injury burden, while occupational disease claims remained limited in all sectors. All recorded PAK cases in manufacturing, mining, and palm oil were classified under non-specific "lain-lain" coding categories, whereas fisheries and telematics reported no occupational disease claims during the study period. Random Forest analysis identified diagnosis specificity, JKK–PAK ratio imbalance, and sectoral characteristics as the variables most strongly associated with potential under-reporting signals. K-Means clustering further categorized sectors into distinct surveillance typologies, distinguishing sectors characterized by minimal occupational disease visibility and limited diagnosis diversity from sectors with comparatively lower occupational risk profiles. The study suggests that potential under-reporting within Indonesia's occupational surveillance system may involve not only missing cases but also limited diagnostic specificity and inconsistent disease attribution. The integration of machine learning analytics and administrative surveillance data demonstrates the potential of data-driven approaches for identifying hidden reporting inconsistencies and supporting more prevention-oriented occupational health governance.
Unknown authors· International Journal of Dat...· 0 citations
This paper assesses the value-add of AI in safety performance by deeply integrating advanced industrial safety engineering risk analysis methodologies, including Hazard Identification and Risk Assessment (HIRA), Fault Tree Analysis (FTA), and Failure Mode and Effects Analysis (FMEA).
Shruti Pawar and Dr Neeta Banger· International Journal of Adv...· 0 citations
Construction safety has remained a major project risk because workers perform their activities in dynamic, temporary and equipment-intensive surroundings where the nature of hazards may change within a very short time. This paper reviews machine-learning-based worker-safety prediction in construction project environments and develops the reviewed findings into a practical risk-control framework. Using the verified coded Excel dataset as the authoritative evidence source, a PRISMA-informed systematic literature review was conducted. The final synthesis included 79 primary empirical or technical studies published during 2016–2026, while 10 background reviews and seven excluded or reclassified records remained outside the primary evidence base. The selected studies were coded according to bibliographic characteristics, AI/ML technique, specific algorithm, data source, modality, dataset, prediction or detection target, safety domain, outcome, availability of performance metrics, explainability method, major findings, safety implications, risk-control relevance and verification notes. The mapping findings show a clear concentration on computer vision and deep learning, particularly for PPE detection, monitoring of safety compliance, site-hazard detection, fall-related risk and worker–equipment interaction. However, traditional machine learning continues to be important for structured accident, injury and safety-indicator datasets. NLP, LLMs, wearable sensors, IoT-based analytics and multimodal approaches are also emerging for accident narratives, safety reports, physiological risk and contextual safety reasoning. The major limitations are related to data quality, class imbalance, small or customised datasets, weak cross-site validation, interpretability, privacy, workflow integration and incomplete extraction of full-text model-performance results. Since 73 included studies still need manual extraction of full-text metrics, this study is positioned as systematic mapping, thematic synthesis and framework development rather than a statistical meta-analysis. Therefore, the main contribution is a worker-safety prediction and risk-control framework that establishes a relationship among data acquisition, preprocessing, modelling, explainability, managerial decision-making, intervention and continuous learning.
Manzur Ashraf, Humayra Ali, Muhammad Mohiul Islam· Australian Journal of Artifi...· 0 citations
Job Safety Analysis (JSA) and pre-task planning can benefit from prior incident records, yet historical accident data is often stored as unstructured narratives that are difficult to consult at the point of planning. A novel framework centered on large language models (LLMs) for highway construction safety reporting and planning is proposed as a foundation for future agentic applications, prioritizing deterministic, local inferencing. The first aim is to enable classification and quality scoring of incident narratives for existing and future reporting purposes. The second is to evaluate retrieval of relevant historical accidents, related imagery, and trusted industry documents for incorporation into daily safety plans. Neural probes were trained to classify incidents along four multiclass and two binary Occupational Injury and Illness Classification System (OIICS) fields and to derive an overall quality score, evaluated on a test set of over 15,000 narratives and a held-out set of 100 author-labeled records, benchmarked against a majority-vote LLM ensemble. The retrieval of historical accidents, reference imagery, and industry documents was benchmarked across embedding models using standard information retrieval metrics. OIICS classification reached 75% held-out accuracy, though the two binary flags were degenerate. The quality score, while meaningful on one database, was distorted on out-of-distribution fatalities in the held-out dataset. Accident retrieval recovered relevant incidents far above chance, performing best on lexically distinct construction activities. On document question answering, an open-weight decoder embedding model surpassed proprietary models. Overall, this work provides a new framework rooted in local inferencing and text embedding models for future agentic applications, with emphasis on bridging external data to JSA reports.
M. Smetana, Trevor Neece, Lev Khazanovich· 0 citations
Artificial intelligence, when responsibly implemented, represents a transformative adjunct to traditional safety practices – capable of significantly improving construction site safety performance globally – but it must be deployed in tandem with organizational commitment, worker training, and robust safety cultures.
Musaed M. Al-Thubaiti, Saeed S. Al-Shahrani, Ryan A. Alsaihaty· World Journal of Advanced En...· 0 citations
This study formally evaluates the effectiveness of AI models over multiple iterations of the models’ architecture for the domain-specific application of automated construction hazard assessment from multimodal inputs and introduces and validates high-fidelity, game engine-based synthetic images as a solution.
Trevor Neece, A. Fascetti· Journal of computing in civi...· 1 citation· ⚡1
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
