Office and Administrative Support · SOC 43-5041
Meter Readers, Utilities
Read meter and record consumption of electricity, gas, water, or steam.
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 15 tasks O*NET lists for this occupation:
Could an LLM do it?
8%–31%
of tasks: an LLM alone could do the task or cut its time by half. Human raters say 8%, GPT-4 says 31%.
Counting tasks that need extra software around the model: 23%–62%.
Is it happening?
0 of 15 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?
6 of 15 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 |
|---|---|---|
| Answer customers' questions about services and charges, or direct them to customer service centers. | all raters | none seen |
| Leave messages to arrange different times to read meters in cases in which meters are not accessible. | all raters | none seen |
| Update client address and meter location information. (supplemental) | all raters | none seen |
| Read electric, gas, water, or steam consumption meters and enter data in route books or hand-held computers. | 2 of 3 raters | none seen |
| Upload into office computers all information collected on hand-held computers during meter rounds, or return route books or hand-held computers to business offices so that data can be compiled. | 2 of 3 raters | none seen |
| Verify readings in cases where consumption appears to be abnormal, and record possible reasons for fluctuations. | 2 of 3 raters | none seen |
| Report to service departments any problems, such as meter irregularities, damaged equipment, or impediments to meter access, including dogs. | 1 of 3 raters | none seen |
| Report lost or broken keys. | 1 of 3 raters | none seen |
| Collect past-due bills. (supplemental) | 1 of 3 raters | none seen |
| Walk or drive vehicles along established routes to take readings of meter dials. | no rater | none seen |
| Install new or replace broken meters. | not rated | n/a |
| Inspect meters for unauthorized connections, defects, and damage, such as broken seals. | no rater | none seen |
| Perform preventative maintenance or minor repairs on meters. | no rater | none seen |
| Dig dirt away from meters to take readings. | not rated | n/a |
| Connect and disconnect utility services at specific locations. (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 Office and Administrative Support
- Customer Service Representatives (70% observed)
- Data Entry Keyers (67% observed)
- Statistical Assistants (51% observed)
- Desktop Publishers (46% observed)
- Secretaries and Administrative Assistants, Except Legal, Medical, and Executive (45% observed)
- Office Clerks, General (45% observed)
- Receptionists and Information Clerks (43% observed)
- Human Resources Assistants, Except Payroll and Timekeeping (40% 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.