Life, Physical, and Social Science · SOC 19-5012
Occupational Health and Safety Technicians
Collect data on work environments for analysis by occupational health and safety specialists. Implement and conduct evaluation of programs designed to limit chemical, physical, biological, and ergonomic risks to workers.
Will AI replace this job?
The data cannot say, and anyone who gives you a percentage for “replaced” is guessing. What it does show is a gap between three things, for the 26 tasks O*NET lists for this occupation:
Could an LLM do it?
12%–23%
of tasks: an LLM alone could do the task or cut its time by half. Human raters say 23%, GPT-4 says 12%.
Counting tasks that need extra software around the model: 42%–69%.
Is it happening?
0 of 26 tasks
show measurable use in Claude conversations (Anthropic Economic Index). Anthropic’s occupation score, which counts only tasks that were also rated feasible, is 0%. Observed is a subset of “could” by construction.
Claude only, mostly US users, sampled periods in 2025. Not all AI use, and not employment.
What is left?
7 of 26 tasks
were rated “no LLM speed-up” by every rater (so, by Anthropic’s method, they cannot show up as observed use either). That is the part of the job these studies see no sign of AI touching.
It does not mean those tasks are safe. It means these methods found nothing there.
Task by task
Sorted with the tasks seen in use first. The three coloured squares are the three raters, in this order: human raters, GPT-4, GPT-4 with an alternative rubric. “Usage score” is Anthropic’s published task_penetration value; 0 means none seen.
| Task (O*NET 31.0) | Could an LLM? | Claude usage score |
|---|---|---|
| Maintain all required environmental records and documentation. | all raters | none seen |
| Provide consultation to organizations or agencies on the workplace application of safety principles, practices, or techniques. | all raters | none seen |
| Recommend corrective measures to be applied based on results of environmental contaminant analyses. | all raters | none seen |
| Prepare or review specifications or orders for the purchase of safety equipment, ensuring that proper features are present and that items conform to health and safety standards. | all raters | none seen |
| Conduct worker studies to determine whether specific instances of disease or illness are job-related. | all raters | none seen |
| Review records or reports concerning laboratory results, staffing, floor plans, fire inspections, or sanitation to gather information for the development or enforcement of safety activities. | all raters | none seen |
| Educate the public about health issues or enforce health legislation to prevent disease, to promote health, or to help people understand health protection procedures and regulations. | all raters | none seen |
| Prepare documents to be used in legal proceedings, testifying in such proceedings when necessary. | all raters | none seen |
| Collect data regarding potential hazards from new equipment or products linked to green practices. | all raters | none seen |
| Maintain logbooks of daily activities, including areas visited or activities performed. | all raters | none seen |
| Plan emergency response drills. | 2 of 3 raters | none seen |
| Examine credentials, licenses, or permits to ensure compliance with licensing requirements. | 2 of 3 raters | none seen |
| Confer with schools, state authorities, or community groups to develop health standards or programs. (supplemental) | 2 of 3 raters | none seen |
| Collect data related to ecological or human health risks at brownfield sites. (supplemental) | 2 of 3 raters | none seen |
| Perform tests to identify any potential hazards related to recycled products used at green building sites. (supplemental) | 2 of 3 raters | none seen |
| Evaluate situations or make determinations when a worker has refused to work on the grounds that danger or potential harm exists. | 1 of 3 raters | none seen |
| Train workers in safety procedures related to green jobs, such as the use of fall protection devices or maintenance of proper ventilation during wind turbine construction. | 1 of 3 raters | none seen |
| Conduct interviews to obtain information or evidence regarding communicable diseases or violations of health or sanitation regulations. (supplemental) | 1 of 3 raters | none seen |
| Examine practices at green building sites to determine whether adherence to green building standards alters risks to workers. (supplemental) | 1 of 3 raters | none seen |
| Supply, operate, or maintain personal protective equipment. | no rater | none seen |
| Test workplaces for environmental hazards, such as exposure to radiation, chemical or biological hazards, or excessive noise. | no rater | none seen |
| Inspect fire suppression systems or portable fire systems to ensure proper working order. | no rater | none seen |
| Verify availability or monitor use of safety equipment, such as hearing protection or respirators. | no rater | none seen |
| Prepare or calibrate equipment used to collect or analyze samples. | no rater | none seen |
| Help direct rescue or firefighting operations in the event of a fire or an explosion. (supplemental) | no rater | none seen |
| Test or balance newly installed HVAC systems to determine whether indoor air quality standards are met. (supplemental) | no rater | none seen |
Try it with your own task list
Sources: task list from O*NET® Database (U.S. Department of Labor, CC BY 4.0; modified by noworkleft.com); “could” labels from Eloundou et al., GPTs are GPTs (OpenAI/University of Pennsylvania, MIT licence, labels from early-2023 models); “seen in use” from the Anthropic Economic Index (Claude usage only, CC BY). Dates, versions and limits.
Other jobs in Life, Physical, and Social Science
- Political Scientists (45% observed)
- Social Science Research Assistants (44% observed)
- Survey Researchers (43% observed)
- Astronomers (38% observed)
- Sociologists (38% observed)
- Chemical Technicians (31% observed)
- Physicists (27% observed)
- Chemists (26% observed)
Data built 2026-11-06. O*NET 31.0. Task match rates: 99.4% with the OpenAI/Penn labels, 99.2% with Anthropic usage.