Why the drill floor still needs a crew
Drilling is a physical job with a thinking job bolted on top. A rotary drill operator starts and stops the rotary table, watches the weight on the bit, and keeps the slush pumps moving mud at the right rate. When the pipe sticks or the hole starts taking fluid, someone has to read the gauges, feel the brake, and decide in seconds. Software can advise on that. It cannot stand there.
The other half of the day is hands and tools. Making and breaking connections, changing bits, greasing and repairing the hoist and pump, rigging up and rigging down a truck-mounted derrick on uneven ground in bad weather. Our estimate of how much of this job is physical rather than screen-based is shown above, and it sits in the mobile-robot tier: machines that would need to move around a worksite, not just run a program. That hardware exists in pilot form, not in the average land crew.
Automated and autonomous drilling systems are real, and they are improving on the directional and optimization side. But a rig is a crew. The operator also directs roughnecks, trains new hands, and signs off on safety checks. So when people ask whether AI will replace rotary drill operators, the honest answer is that the paperwork and the data reading move first, and the floor work moves last.
What software handles, what it assists, and what stays with the operator
Start with the part a model could take on by itself: about 0% of task time. That is the clerical and analytical end — keeping records of drilling progress and mud volumes, and preparing shift and well reports. These are structured, repeatable, and already half-digital on many rigs.
The assist share is 7%. Here the operator keeps the decision and the system sharpens it. Monitoring pressure, torque, and rate-of-penetration readings is one example: sensors stream the data and a model flags a trend earlier than a tired human eye. Choosing drilling parameters to protect the bit is another. The judgment, and the liability, stay with the person on the brake.
The rest, 93%, needs a person on site. Connecting and disconnecting sections of drill pipe, inspecting and repairing pumps, derricks, and hoisting gear, and supervising a crew through a trip out of the hole are not screen tasks. Overall coverage for this occupation reads 6 out of 100 on the question of whether AI can do the work today. You can see how that figure is built on how coverage is measured.
How strong is the evidence?
Weaker than the headlines suggest. The parity grade for this job is D, which means no study has tested an AI system against a qualified rotary drill operator doing this job’s full task mix. So we publish no parity number for it. Industry papers report autonomous drilling performance on specific wells and specific sections, which is useful, but that is not the same as measuring a system against a person across connections, maintenance, crew direction, and emergency response.
What would settle it: a public, repeatable trial on comparable wells, where an autonomous system and a human-led crew run the same program and the results include non-productive time, safety incidents, equipment damage, and how often a person had to take over. Until something like that exists, treat any confident claim about this job — including ours — as an estimate with its reasoning shown. The method is written up in full on the methodology page.
Good to know: an occupation can score well on the headline question and still lose jobs, because drilling employment rises and falls with oil prices and rig counts, not with AI.
When the picture could shift
Most likely after 2048 (8 in 10 of our scenarios). That window is wide on purpose, because hardware timelines are harder to call than software ones. What the range means, and how it is produced, is explained on the replacement-year method.
Two things could pull it earlier. First, remote operations centers: if more of the decision work moves off the rig to a control room, one specialist plus a system may cover several wells, which thins the crew even without a robot on the floor. Second, cost. The software side of this work is cheap to run compared with an hourly wage, and that gap pushes operators to automate monitoring and reporting first.
Two things hold it back. The physical share of the job needs machines that can move, lift, and work safely around high-pressure equipment, and that class of robot is not yet common on land rigs. And the safety case is unforgiving: a failure while tripping pipe or controlling a well is not a software bug, it is a blowout risk. Regulators, insurers, and operators all move slowly on that. The guide to robots and physical jobs covers why the hardware lags the models.
How to stay needed on a rig
Lean into the work that stays human. Troubleshooting a hole that is not behaving — losses, stuck pipe, pressure changes — rewards experience and is the hardest part to hand over. Mechanical repair on the pump, hoist, and derrick keeps you useful on any rig, in any price cycle. And crew leadership, including training new hands and running safety checks, is the route to toolpusher and driller roles.
Two skills worth adding. One: reading and questioning drilling data and automated recommendations, so you can tell when a system is wrong rather than following it. Two: a well-control certification kept current, because the paper you hold decides what you are allowed to run.
If you want to compare nearby work, look at derrick operators, service unit operators, and roustabouts, all in the extraction workers family. Pay and demand context helps too: median pay for this occupation was $67,890 and employment about 12,600, with projected growth of 0.8% over 2025 to 2035 (BLS, 2025). The wider picture for drilling and extraction sits on the mining, oil and gas sector page.
Headline score for this job: 85 out of 100 (higher is safer). Put it beside another job on the compare tool, or see where hands-on roles land on the list of jobs that mostly need a person.