Do occupational health and safety tools that utilize artificial intelligence have a measurable impact on worker injury or illness? Findings from a systematic review
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Background Artificial intelligence (AI) holds promise as a tool that can be used by practitioners in the field of occupational health and safety (OHS). This study aimed to identify AI applications specifically used for OHS and examine their impact on worker morbidity or mortality outcomes. Methods We conducted a comprehensive systematic review. We searched six databases to identify published quantitative studies of OHS AI applications across the hierarchy of controls that were published between years 2018 to 2024. Title/abstract and full-text screening was conducted to identify eligible studies which were then assessed for quality and risk of bias and synthesized. Results Of the 1255 articles identified by our search, only two met eligibility criteria; one of which was appraised as medium quality and the other as low quality. The one medium quality study identified by our review was an AI-based chatbot health promotion tool which was shown to improve musculoskeletal symptoms. Our systematic review shows that we are at the early stages of understanding the role AI can play in OHS and it may be premature to recommend the wide-spread use of AI for health and safety practice within workplaces. Conclusion There is a critical need for future research to unpack how considerations taken in the development and adoption of workplace AI tools for OHS can determine their effectiveness in addressing worker injury or illness. Systematic review registration: PROSPERO CRD42023414422.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.023 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.413 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it