Why the charge still gets set by hand
The honest answer to whether AI will replace explosives workers sits in the score above, and the reason is the shape of the work. Someone has to walk the bench, read the rock face, load each hole, tie in the circuit and clear the area before firing. Software can model a pattern. It cannot push stemming down a hole, feel a loose seam, or call off loading because a face looks unstable this morning.
Two tasks explain most of it. Placing and priming charges in drilled holes is physical, one-off work in changing ground. Examining the blast area afterward for misfires and unexploded material is judgment under risk, with a person accountable for the call. Ordnance handling adds the same problem in sharper form: an item that was never designed to be safe to move.
There is also the paper side. Permits, magazine records, and logs of the type and amount of explosives used all name a qualified, licensed person. Accountability does not transfer to a model. That keeps a trained human in the loop even where a machine does the lifting.
What AI runs, what it assists, and what people keep
The share of task time that still needs a person here is 91%. Across all tasks, the share AI can handle today reads 5 out of 100 on our coverage measure; the method behind that figure is set out in how coverage is scored.
The slice AI could finish on its own, 0%, is clerical. Think of writing up a shot report from recorded inputs, or keeping the running record of charges drawn from the magazine against what was fired. Both are structured and repeated, and both are already partly software.
The assisted slice, 9%, is where the interesting change is. Blast design, delay timing, and predictions of ground vibration and fly-rock are modeling problems. Drone survey and photogrammetry give better face profiles than a tape and a notebook. Sensors on holes and seismographs around the site give feedback a blaster can learn from. None of that loads a hole.
What stays with people is the field half: loading and priming, inspecting the shot area, deciding when conditions are wrong, and dealing with a misfire. For a sense of how physical that is, the page shows most of this job’s task time as bodily work, and the machine class it would take is mobile robots that can travel rough, unmade ground and manipulate sensitive items. That tier exists in research and in some specialist disposal units; it is not a hiring substitute for a crew.
What has actually been tested
Not much, and that matters. The evidence grade for the quality question on this job is D, which on our scale means there is no direct test of a model or a machine against a qualified blaster. So no parity number is published here, and anyone quoting one for this job is guessing. You can read what each grade requires on the quality parity method page.
What would settle it: a published trial of remote or robotic loading and firing on production benches, measured against crews on the same ground; or an audited record of autonomous ordnance disposal compared with trained technicians on the same class of item. Until something like that exists, the useful evidence is the labor data. The US Bureau of Labor Statistics counts roughly 5,100 people in this occupation, with median pay of $61,390 a year and projected employment change of about 0% from 2025 to 2035 (BLS, 2025). That is a small, flat, licensed trade, not a shrinking clerical pool.
Good to know: a flat projection means little replacement pressure, but also few openings, so entry usually comes through helper and driller roles rather than straight in.
When the timing could move
Most likely after 2048 (8 in 10 of our scenarios). What that window measures, and why it is a range rather than a date, is explained on the replacement year method page.
Two things could pull it earlier. Remote-controlled and semi-autonomous loading rigs are already used in underground mining for the most dangerous faces, and safety rules tend to push adoption where a person would otherwise stand in the blast zone. Falling hardware costs help too: the running cost of the software side of this work is a fraction of a crew’s monthly cost, so operators have a reason to automate anything a machine can reach.
Two things hold it back. Ground conditions change hole by hole, and a rig that works in a neat underground drift is not ready for a weathered quarry face or a mixed demolition site. And the regulatory frame assigns licensing, storage and firing responsibility to named, qualified people. Changing that is slower than building the robot. Other hands-on trades sit in a similar spot, which is the pattern traced in our guide to robots and physical jobs.
How to stay needed in blasting work
Lean into the three parts no system is close to owning. First, shot inspection and misfire handling, including the decision to stop. Second, loading and priming in awkward or degraded ground, where the plan on the screen meets the rock. Third, ordnance identification and disposal, where experience with a specific item class is the whole job.
Two skills raise your value fast. One is reading blast design and vibration modeling output well enough to challenge it, so you are the person who signs off rather than the person who is handed a pattern. The other is compliance craft: magazine management, permits and clean records, which is the part regulators and insurers care about most.
If you are weighing adjacent moves, close work sits with earth drillers, continuous mining machine operators and hazardous materials removal workers. The wider picture for the trade is on the extraction workers family page and in the mining, oil and gas sector.
Next step: put this job beside a neighboring one on our side-by-side comparison, see where it lands among jobs that mostly need a person (our top band, Nah.), or read how the scoring works.