Why the rock keeps this work with people
Roof bolting is the step that stops a tunnel from coming down. After a cutting machine takes the face, bolters drill into freshly exposed roof and set bolts that tie weak rock to stronger ground above it. The space is low, dark, dusty and different every few feet. That is the short answer to the question people type as will ai replace roof bolters: the job is one long judgment call made inside a hazard.
Two tasks carry most of that judgment. The first is sounding and testing the roof: tapping, listening, watching for drummy rock, cracks and water, then deciding where supports go before anyone works under it. The second is scaling and barring down loose rock by hand, which depends on feel and on knowing what the last cut did to the ground. Both change with the seam, the stress, the moisture and the shift before yours.
Size matters too. The Bureau of Labor Statistics counts roughly 2,160 roof bolters in the United States, with median pay of $78,540 a year (BLS, 2025). BLS also projects employment falling about 18.9% between 2025 and 2035 (BLS, 2025). That drop tracks demand for coal and ore, not software. A trade this small gives machine builders little reason to fund a dedicated underground bolting robot, even where the engineering is possible.
What software does, what it assists, and what stays hands-on
Nothing on this job’s task list sits in the group AI handles on its own today. The share of task time software can take without a person is 0%. Office tools can write a report; they cannot place a bolt in a roof they cannot reach.
The assist group is also empty on our list so far, at 0% of task time. In practice, digital help sits beside the work rather than inside it: ground-control plans, monitoring data and training materials. The bolting cycle itself is still performed, not reviewed.
That leaves everything with the crew, at 100% of task time. Positioning the bolting machine and its safety jacks under an unsupported roof, drilling holes to the depth the plan calls for, setting resin or mechanical bolts, checking torque, and watching the face while a partner works are all physical, in-place tasks. Our coverage figure, which answers can AI do it, sits at 1/100; the coverage method explains how that share is built.
How strong the evidence is
There is no direct test of AI against roof bolters in this job yet. Our quality parity grade is D, which is the grade we use when no study has measured machine performance against a qualified person doing this work. For that reason we publish no parity number for bolting, and you should treat any outside claim of one with care.
What would settle it is specific: a trial of automated or tele-remote bolting rigs in production headings, reporting bolts installed per shift, hole depth and torque accuracy, roof-fall and injury rates, and how often a person had to enter the unsupported area to fix the machine. Until somebody publishes that, the honest position is unmeasured. How we grade evidence is set out in the quality parity method.
What could move the date, and what holds it
Most likely after 2048 (8 in 10 of our scenarios). The replacement-year method explains what that window measures and how the spread is produced.
Two things could pull it earlier. Tele-remote and semi-automated bolting rigs already exist in hard-rock mining, and each step that moves an operator out of the unsupported heading removes a reason to keep a person at the face. Safety regulation can push the same way, because a rule that bans entry under unsupported roof makes remote equipment the cheaper path.
Two things hold it back. The physical share of this job is total, and the robotics tier it needs is dexterous and humanoid-like: reaching, bracing, feeling resin set, clearing a jammed drill steel in a four-foot seam. Capital is the other brake. Underground fleets are bought to last, mines are shrinking rather than expanding, and a small workforce spreads development cost thinly. The cost panel above shows how the machine and human numbers compare on this job.
Good to know: automation underground usually arrives as remote control first, with the operator moved back down the entry rather than removed from the mine.
How to stay needed at the face
Lean into the parts of the job that carry responsibility for other people. Ground assessment comes first: reading roof conditions, sounding, and calling when a plan needs to change. Second, temporary support and scaling, where sequence and timing decide whether anyone gets hurt. Third, pre-operational checks and fault finding on the bolter itself, since a crew that can diagnose hydraulics and drill feed keeps the section moving.
Two skills raise your floor. One is ground control and mine safety certification, including the paperwork and inspection side that supervisors need. The other is equipment electronics and remote operation, because the people who run and troubleshoot tele-remote rigs are the ones mines keep when a fleet is modernized.
Close work is worth comparing. Look at continuous mining machine operators, underground loading and moving machine operators, and extraction worker helpers, all in the extraction workers family. You can put any two of them next to each other on the job comparison tool, see the wider picture on the mining, oil and gas sector page, or read where hands-on trades land on the list of jobs least exposed to AI. The scoring itself is open: see how we score occupations.