EXPOSED
This is a catch-all bucket for hands-on production work that doesn't fit a named title — feeding and tending machines, hand-assembling parts, inspecting output, packing, moving material between stations. Language AI can't do any of that, which is real protection; the threat is industrial automation and vision-based inspection, and factory floors are exactly the controlled environments where robotics actually works. There is no license, no client relationship, and little discretion, so when a line is re-tooled the job is designed out rather than augmented.
Dipped in 2020, then grew past where it started.
Median pay $29,800 → $40,110 +7.7% 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
+0.5%
Percentage only. The projection counts a different population from the 251,700 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Exposed, but growing
AI can already do a lot of these tasks, and the BLS still expects +0.5% more of these jobs by 2034. Demand for the output is growing faster than the work is being automated away — the mechanism BLS gives for software developers, and the combination people most often misread as an error.
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.
~31,600 openings a year on average, including replacing people who leave.
HandBaterBluerBonerCaserCurerDaterDoperFinerGluerLacerLayerLinerMaterMixerOilerPagerPolerRakerRiserRulerTuberWiperBagger
The BLS uses Production Workers, All Other for work that doesn't fit any named occupation, so it covers roles that have little in common with each other. Two consequences worth knowing before you read anything below:
Holding it up: embodiment . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier At 12 rather than 16, the physical steps — loading fixtures, deburring by hand, hand-packing odd-shaped product — genuinely can't be typed into a prompt, but they sit inside standardized cycle times on a repeatable line, which is precisely the work robotic cells and vision inspection have already taken over in higher-volume plants.
Hands-on in uncontrolled environments A 13 reflects that the whole shift happens on your feet at a machine or bench with material in your hands, but the floor is a lit, flat, climate-managed indoor space with fixed station layouts — not a roof, a trench, or a customer's basement, which is what pushes a score into the high teens.
No licence, no signature requirement A 2 is honest: no state license gates entry, the machine guarding and OSHA duties rest on the employer under 29 CFR 1910, and QC sign-off traces to the quality department or the engineer who wrote the spec, so nothing legally requires that a human hold this station.
Executes defined procedures on defined inputs A 5 fits work bounded by the work order, the travel sheet, and go/no-go gauges, where the discretion you do hold — pulling a suspect batch, calling maintenance when a feeder jams — is real but escalates upward within minutes rather than resolving on your authority.
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 (12/20 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 (2/20) is whether the law requires a licensed human to sign. Trust premium (3/20) is whether buyers specifically pay for a person. Judgment and accountability (5/20) 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 10 of this occupation's 35 points (29%).
Embodiment (13/20) 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 50/100, still EXPOSED.
Task-mix shift within the bucket: this SOC genuinely has two tiers. If repetitive feed/pack/count work is automated first, what remains is changeover, fixture setup, jam and exception recovery, and hand-finishing of odd geometries — work that resists cell-level automation because volumes don't amortize tooling. Watch high-mix low-volume reshoring (CHIPS-adjacent suppliers, defense small-lot) where lot sizes stay under automation payback.
Regulated-sector production carve-outs where a named human must sign a batch record: FDA 21 CFR 211/820 device and drug lot-release, FSMA preventive-control monitoring records, and AS9100/Nadcap special-process operator certification. If FDA's evolving guidance on AI in manufacturing (the 2025 draft on AI credibility in regulated decisions) settles on requiring an identified, trained human to attest to automated inspection results rather than validating the system alone, the named-verifier role inside 51-9199 hardens.
Formal designation of line operators as the 'qualified person' in quality systems — signing deviation/nonconformance reports, authorizing stop-line and hold decisions when vision inspection flags ambiguity. Already contractual in some UAW/USW quality-agreement language and in ISO 9001 competence requirements; expands if defect-recall liability pushes firms to name a human decision owner for every disposition call.
Growth of the sub-bucket doing non-repeatable physical work: rework, teardown/refurbish, and remanufacturing lines where each incoming unit differs. Right-to-repair statutes (Minnesota 2023, Oregon 2024) and EU battery/ecodesign remanufacture rules expand exactly this unstructured-input physical work, which is much harder to fixture than greenfield assembly.
Narrow and mostly not available: certified-handmade or origin-labeled goods (small-batch food, instruments, artisanal consumer products) carry a human-made premium, but they are a tiny share of the 251,700 and the premium attaches to the brand rather than the worker. No realistic route to a broad trust premium in commodity production.
The limit. The structural ceiling is that no license attaches to the occupation itself — the liability shields above protect a designated role a firm assigns, not a credential a worker carries, so employers can concentrate sign-off authority in a handful of quality technicians and design the rest of the headcount out. Regulated-sector protection also only covers the pharma/device/food/aerospace slice; the general-manufacturing majority has no analogous mechanism.
| Los Angeles-Long Beach-Anaheim, CA | 11,190 | $39,520 -1% |
| Atlanta-Sandy Springs-Roswell, GA | 10,720 | $40,970 +2% |
| Dallas-Fort Worth-Arlington, TX | 6,150 | $39,540 -1% |
| Chicago-Naperville-Elgin, IL-IN | 4,810 | $41,250 +3% |
| Nashville-Davidson--Murfreesboro--Franklin, TN | 4,790 | $40,510 +1% |
| New York-Newark-Jersey City, NY-NJ | 4,740 | $40,610 +1% |
| Charlotte-Concord-Gastonia, NC-SC | 3,710 | $42,620 +6% |
| St. Louis, MO-IL | 3,530 | $40,900 +2% |
| Lewiston, ID-WA | 40 | $93,420 +133% |
| Ames, IA | 70 | $70,690 +76% |
| Hanford-Corcoran, CA | 160 | $69,970 +74% |
General Motors · Hyundai · Ford · BMW · Kia · Toyota · Rockwell Automation
DW News reports that thousands of factory workers in India are wearing body-mounted cameras to capture egocentric footage of their manual tasks, which is used as training data for AI-powered humanoid robots; no specific employer is named.
Kia's Gwangju plant was reported to have deployed AI for industrial safety management, receiving an award from South Korea's Ministry of Employment and Labor.
CIO.com reports that Ford has rehired 350 former employees after disappointing results from AI systems that had replaced their work.
People Matters reports GM has introduced 50 collaborative robots at its Factory Zero plant, affecting more than 1,000 workers.
The Guardian reports that Indian factory workers are being asked to record video of themselves performing their tasks to generate training data for AI and robotics systems, with workers raising concerns about being replaced.
WION reports that General Motors has introduced 50 robots at its Factory Zero plant while over 1,000 jobs were cut.
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.