Why loading and offbearing stays with people
This job happens inches from moving equipment. Machine feeders and offbearers lift stock into a press, saw, oven or packaging line, then pull the finished product off the other end and set it aside for the next step. The materials rarely cooperate. Sheet, board, bags, trimmings and odd-shaped parts all behave differently, and a gripper tuned for one part is useless on the next one without new tooling.
The other reason is trouble. When stock binds or a part hangs up, somebody has to reach in, read what went wrong and free the jam without damaging the tool or the batch. The same person notices the smell of a hot bearing, the change in the sound of a feed roll, or the sheet that went in crooked. That mix of judgment and hands is why coverage sits at 6 out of 100 on the question of whether AI can do the work today. Coverage measures the share of task time software can handle, and it is explained on the coverage method page.
None of that means the job is standing still. The honest story here is task erosion and fewer new hires rather than the role disappearing. The Bureau of Labor Statistics counts about 42,330 of these jobs in the US, with median pay of $41,220, and projects employment falling 13.1% between 2025 and 2035 (BLS, 2025). The pressure comes from factories buying feeders, robot cells and conveyors, not from a chatbot.
What software does, what it assists, and what it leaves to the floor
The tasks AI can take outright are the paperwork ones. Logging production counts, recording quantities and material types, and passing shift data up to a scheduling system are all things a sensor and a database do faster than a clipboard. That slice of task time is 0% of the job.
Assistance is the more interesting band, at 9% of task time. Watching machine operation for faults is now shared work: vibration and temperature sensors flag a problem before a person would hear it, but the person still decides whether to stop the line. Inspecting output for defects is similar. Vision systems sort obvious rejects at speed, while borderline parts and new defect types go back to the operator.
What stays with people is the bulk of it, 91% of task time. Feeding and offbearing material by hand tops that list, along with clearing jams and freeing stuck stock, cleaning equipment and the work area, and signaling coworkers to start, slow or stop a line. These are short, varied, physical tasks in a space built for people, and that is exactly the kind of work that machines handle last.
What the evidence actually shows
There is no direct test of AI against people in this occupation yet. The quality-parity grade is D, which is our marker for not measured, so we publish no parity number for machine feeders and offbearers. The grading scale is set out on the quality parity method page.
What would settle it is plant-floor measurement rather than a lab demo: timed trials of a robot cell against a human feeder on mixed stock, jam-recovery rates on a live line, and cost and uptime figures over a full year rather than a pilot week. Until that exists, the score leans on the task mix and on what the equipment can physically do. You can read how all three questions fit together in the scoring methodology.
When this could change
Most likely after 2044 (8 in 10 of our scenarios). For what that window does and does not mean, see the replacement-year method.
Two things could pull it earlier. The first is cheaper, more general handling hardware, since the robotics profile for this job sits in the mobile-robot tier rather than the fixed-arm tier, and mobile units keep getting less expensive. The second is plant redesign: when a line is rebuilt, feeding and offbearing are often designed out before anyone buys a robot.
Two things hold it back. Capital cost and payback come first, because a new cell has to beat a wage that the BLS puts near $41,220 a year (BLS, 2025), and short runs make that math hard. Second is variability. Small batches, changing stock sizes, dusty or wet conditions and tight safety rules around moving equipment all slow installation, and every jam a robot cannot clear brings a person back to the line.
How to stay needed
Lean into the parts of the work that machines keep handing back. Jam clearing and recovery is the big one, because it is the task that decides whether an automated cell runs unattended or not. Setup and changeover work, including weighing and positioning stock for a new run, is the second. Keeping equipment and the work area clean and safe is the third, and it is also how you learn which machines are about to fail.
Two skills matter more than the rest. Machine tending at a higher level, meaning basic troubleshooting, minor adjustment and knowing when to call maintenance, moves you from feeding the machine to keeping it running. Reading the data the line now produces, so you can act on a sensor alert instead of waiting for a supervisor, is the other.
What to do: ask your supervisor to train you on the cell or conveyor that is most likely to be automated next, rather than the one you already know best.
Nearby jobs share most of these tasks and face similar pressure. The closest are packers and packagers, hand, laborers and freight, stock and material movers, and conveyor operators and tenders. You can put any two side by side on the job comparison tool, or look at the wider material moving workers family to see where the task mix differs.
For the plant-wide picture, the manufacturing sector page collects the jobs on the same floor, and our guide to humanoid robots and physical work covers what the hardware can and cannot do yet. If you want a view across every scored job, the full job rankings are searchable, and the Still needs a human figure for this role is 85 out of 100 (higher is safer).