Why turning work stays at the machine
The job starts before the spindle turns. Someone reads the print, checks the stock, builds the workholding, trues up a chuck or collet, sets tool offsets and cuts a first piece. None of that is typing. It is hands on metal, with a micrometer in one hand and a tool holder in the other.
Then comes the part that decides whether the run is good: watching and listening. Chatter in a long part, a chip color that says the insert is done, a finish that feels wrong under a thumbnail. Operators adjust feeds and speeds, swap an insert, shim a jaw and keep the part inside tolerance. Software can suggest a change. It cannot clear a bird’s nest of swarf or re-indicate a slipping part.
Shops also make the problem harder than a demo. Short runs, mixed materials, old machines beside new ones, a bar that arrived slightly bent. Most of this occupation’s task time is physical work in that setting, which is why turning still sits with people even as the digital parts of the job move. You can see the same pattern across metal and plastic workers and across manufacturing as a whole.
What software runs, what it assists, and what people keep
The tasks AI can handle on its own account for 0% of task time here. These are the paperwork-shaped pieces: pulling dimensions and tolerances out of a drawing or model, and logging production counts, scrap and machine hours. Both are text and numbers, so they move first. Our coverage measure explains what counts as AI doing a task rather than helping with it.
Assisted work covers 22%. Programming is the clearest case: a model can draft a turning program or a tool list, and a machinist checks the approach, the clearances and the order of operations. Condition monitoring is the second: spindle load and vibration data can flag a worn insert or a bearing going bad before a part scraps, but someone still decides whether to stop the run.
People hold 78%. That is setup and alignment, loading and clamping awkward stock, in-process measuring with micrometers, gauges and indicators, tool changes, deburring and the judgment calls when a part drifts. A robot can tend a machine that is already set up and repeating. Getting the machine to that point is the skilled part.
What the evidence shows so far
No study has tested an AI system against a qualified turning operator on this job’s real tasks. Our evidence grade reflects that: D on an A to D scale, where D means not measured, so we publish no parity number. A fair test would need a shop floor, not a benchmark: mixed jobs, unfamiliar prints, worn tooling, and a scored comparison on setup time, first-part accuracy and scrap.
Labor market data is firmer. The Bureau of Labor Statistics counted about 16,710 of these operators in the United States, with median pay near $50,620 a year, and projects employment falling about 11.3% between 2025 and 2035 (BLS, 2025). That decline is mostly work consolidating into CNC roles and fewer entry-level tenders being hired, not turning disappearing. The jobs expected to shrink list shows where this sits against other occupations, and the full method explains how we combine task data with evidence.
When the picture could change
Most likely after 2046 (8 in 10 of our scenarios). Our replacement-year page sets out what that window does and does not claim.
Two things could pull it earlier. Cheap, reliable machine tending with mobile robots would cover the loading and unloading that currently keeps someone at the machine all shift. And software costs for the digital tasks are a fraction of a monthly wage, as the cost panel on this page shows, so the programming and paperwork side has every reason to move fast.
Two things hold it back. The physical share of the work needs grippers, fixtures and sensing that still struggle with chips, coolant and non-rigid parts. And capital is the brake in small shops: a job shop running fifty-part batches rarely justifies a cell that pays off only on long runs. Robot tending also assumes consistent raw stock, which plenty of shops do not have.
How to stay needed
Lean into the work that stays human. Setup and alignment on unfamiliar parts, in-process inspection with hand gauges and a feel for what the print really demands, and problem solving when a part moves, chatters or finishes badly. Those three decide whether a shop can take tricky jobs at all.
Two skills pay off: reading and editing CNC code so you can check and fix what software drafts, and metrology, including CMM work and documented first-article inspection. Both raise what you are worth in the same building.
What to do: ask to run the shop’s setup and inspection sign-off on short-run jobs, since that is the part automation reaches last.
Nearby work is worth comparing before you retrain. Machinists cover a wider range of operations and prints. Computer numerically controlled tool operators sit closer to the programming side. Milling and planing machine setters share most of the same setup skills. You can put any two of them side by side on the compare tool, or look up your own role in the rankings.