EXPOSED
The modal worker splits time between rough field settings — collecting rock, soil and core samples, running downhole and surface instruments, mudlogging at active drill sites — and screen work compiling logs, plotting data and drafting reports, which is the automatable half. Automated core scanners, real-time telemetry, and ML lithology/log interpretation are already eating the description and data-reduction tasks, while the sample handling and rig-side presence stay human because there is no robot that will hike a claim block or babysit a drill floor. No license shields the role and geologists sign the deliverables, so the protection is physical, not legal.
Mixed — a routine tier and a judgment tier. Sample prep, core description, log plotting and report compilation — roughly half the shift — are already being handled by hyperspectral core scanners and ML log-interpretation packages, but rig-side mudlogging, sample chain-of-custody, geophysical instrument setup and troubleshooting a stuck downhole tool keep it out of the single-digit band.
Hands-on in uncontrolled environments. The job happens on drill floors, claim blocks, quarry benches and offshore rigs with H2S monitors and hardhats, hauling core boxes and running trenching and sampling gear in weather nobody controls; it sits at 15 rather than 18 because a real share of hours are spent in a core shack or trailer at a screen.
No licence, no signature requirement. There is no license or registration to hold — the Professional Geologist stamp on the report belongs to the geologist you report to, and the 3 reflects only site-specific tickets like MSHA Part 46/48, HAZWOPER or offshore safety training that gate access to the location, not to the work itself.
Some relationship component. Clients and regulators buy the assay results and the log, not you, though repeat crews and consultancies do keep techs who know a particular basin's stratigraphy or a specific rig's quirks and who the drillers will actually listen to.
Meaningful discretion. You make real calls on the fly — where to cut the sample interval, whether a gas show is worth flagging to the mudlogger, whether an instrument reading is a formation or a bad calibration — but the interpretation, the well decision and the resource estimate belong to the supervising geologist, which caps this at 8.
Has AI actually changed your work?