EXPOSED
The core of the job — setting up machines, dialing in fixtures and workholding, changing tools, hearing a bad cut, deburring, and measuring parts with mics and CMMs — is hands-on work in a shop where every job differs, and today's robotics can't touch most of it. What AI does erode is the programming and planning layer: G-code generation, CAM toolpath selection, feeds-and-speeds lookup, and blueprint-to-setup-sheet translation are increasingly automated, and high-volume run-tending has been shrinking for decades. The modal machinist in a job shop or small-lot manufacturer keeps working; the pure production operator and the programmer-only role are the squeezed tiers.
Part 2020 shock, part continued decline in the years since.
Median pay $44,420 → $58,750 +5.8% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
0% 299,500 → 299,600 on the projections basis
Shrinking, but not obviously because of AI
The BLS projects 0% by 2034, but at 52/100 this work is only moderately exposed — not the profile of a job current AI can simply do. Occupations shrink for many reasons, and the score does not point at automation as this one's cause.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~29,500 openings a year on average, including replacing people who leave.
JobberMachinistMechanistFixture MakerGear MachinistMachine FitterTool MachinistLathe MachinistMetal MachinistCarbide OperatorInstrument MakerManual MachinistOutside MachinistThermometer MakerAircraft MachinistToolroom MachinistFour Slide OperatorLight Fixture MakerPrecision MachinistTool Room MachinistDevelopment MechanicElectrical MachinistFluid Power MechanicProduction Machinist
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation Indicating a casting true within .0005 on a four-jaw, shimming a soft-jaw to hold a thin-wall part without crushing it, and re-cutting a boring bar's approach after chatter shows up are all decisions made by touch and sound at the machine, which is why this sits at 14 rather than 18 — the CAM programming, tool-library selection, and setup-sheet paperwork that used to eat hours of your shift are the parts genuinely being handed to software.
Hands-on in uncontrolled environments You are inside a machine envelope with a dial indicator, hauling 60-lb vises and chucks, dealing with hot chips, flood coolant on the floor, and tramming a head by feel — a 17 rather than 20 because the shop floor is at least a fixed, lit, indoor space, not a trench or a roofline.
No licence, no signature requirement No state licence gates running a Bridgeport or a Haas; NIMS credentials and a journeyman card help you get hired and paid, but if a part fails in the field the liability runs to the shop's ISO/AS9100 quality system and the engineer who signed the print, not to you.
Meaningful discretion You decide whether an out-of-tolerance feature gets scrapped, reworked, or written up for deviation, when to stop a run because tool wear is trending, and how to interpret a print that GD&Ts a datum scheme the fixture can't reach — real calls with money attached, held to 11 because engineering owns the design intent and QC owns the final accept/reject.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (14/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (3/20) is whether the law requires a licensed human to sign. Trust premium (7/20) is whether buyers specifically pay for a person. Judgment and accountability (11/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 21 of this occupation's 52 points (40%).
Embodiment (17/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 64/100, still EXPOSED.
Task-mix shift: as CAM/G-code generation and setup-sheet translation get automated, the surviving job is the judgment tier — first-article inspection, fixture design for one-off geometry, diagnosing chatter/tool deflection, salvaging out-of-tolerance parts, and hard-to-machine alloys (Inconel, titanium) in aerospace/defense small lots. This is a genuine two-tier occupation and the routine tier is already leaving.
AS9100/NADCAP and ITAR-driven customer flowdowns that require a named, qualified human to sign first-article inspection reports (AS9102) and material certs for flight-critical parts, extended explicitly to bar AI-generated inspection sign-off. Similar language is appearing in defense supplier quality clauses and could be tightened by DCMA or prime-contractor audits.
If traceability regimes (aerospace, medical implant, nuclear per ASME NQA-1) push the named machinist/setup tech as the accountable signer on scrap/rework and deviation dispositions rather than a QA department, the role owns consequential ambiguous calls directly.
Narrow route only: prototype and R&D job shops where customers pay for a specific machinist's problem-solving on unmanufacturable drawings, and defense reshoring clauses that pay a premium for domestic human-staffed shops. This is a small share of the 287k and will not lift the occupation broadly.
The limit. Embodiment is already near ceiling at 17 and cannot meaningfully rise. There is no licensure for machinists — no state board, no personal liability — so liability_shield can only rise via customer/audit contract flowdown, which caps it well below a professional-license score. Realistic ceiling around 62-65.
| Chicago-Naperville-Elgin, IL-IN | 11,270 | $59,520 +1% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 9,420 | $62,070 +6% |
| Houston-Pasadena-The Woodlands, TX | 8,140 | $61,280 +4% |
| Los Angeles-Long Beach-Anaheim, CA | 7,850 | $55,660 -5% |
| Detroit-Warren-Dearborn, MI | 7,680 | $59,310 +1% |
| St. Louis, MO-IL | 5,620 | $62,210 +6% |
| Boston-Cambridge-Newton, MA-NH | 4,870 | $72,060 +23% |
| Dallas-Fort Worth-Arlington, TX | 4,660 | $59,830 +2% |
| Urban Honolulu, HI | 250 | $84,410 +44% |
| Florence-Muscle Shoals, AL | 160 | $82,690 +41% |
| Bremerton-Silverdale-Port Orchard, WA | 370 | $80,620 +37% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 52. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.