Why this work stays in the rows
Will AI replace farmworkers? Not in the way the headlines suggest. The heart of the job is hand work in fields and under glass: picking fruit and vegetables when they ripen, transplanting seedlings, hoeing and pulling weeds between plants, pruning, and loading crates onto trucks. Software can plan and watch. It still cannot judge ripeness by touch, reach into a vine without bruising fruit, or carry a full bin over an uneven furrow.
Variety is the other obstacle. A strawberry row, an apple block, and a greenhouse bench of bedding plants are three different problems. Plants grow at different heights, hide fruit under leaves, and change shape week to week. A machine tuned for one trellis system often fails on the next farm over. Add mud, dust, heat, rain, and short harvest windows, and the engineering bar rises fast. Most of the crops that need the most labor are also the easiest to damage.
Money matters too. About 265,500 people hold this job in the United States, with median pay of $35,660 a year (BLS, 2025). The Bureau of Labor Statistics projects employment in the occupation to shrink slightly over 2025 to 2035 (BLS, 2025) — a drift, not a cliff. Harvest machinery is a capital purchase with a short seasonal payback, so growers buy it where one machine covers many acres of a single crop. You can see the same pattern across the agriculture sector, where equipment spreads crop by crop rather than all at once.
What AI does, what it helps with, and what people keep
The tasks our data assigns to AI alone are the paperwork ones: logging what was planted and picked, tracking yields and hours, and turning sensor readings into a plan for the week. That group accounts for 4% of task time on this job, which is why the coverage figure here sits low. For how that figure is built from task time, see how coverage is measured.
The assisted group is bigger and growing. Drone and camera scouting flags weeds, pest damage, and dry patches before a crew walks the block. Sensor-driven irrigation decides when to water a greenhouse bench. Grading lines use machine vision to sort fruit by size and color while people handle packing and the odd items. Tasks AI can help with make up 11% of the work.
What is left is the hand work, and it is the bulk of the day: 85% of task time. Hand-harvesting delicate crops, setting and repairing irrigation lines and fences, thinning and pruning, and moving stock around a nursery all sit here. Our robotics read puts the physical demand of this job in the dexterous humanoid tier, meaning the hardware would need hands and footing close to a person’s before it could take the task outright.
What has actually been tested
There is no direct, published test of AI against people doing this job. That is why the quality-parity evidence grade is D, and why we publish no parity number for farmworkers. The honest position is that nobody has measured a machine crew against a human crew on the same crop, on the same days, at the same quality standard.
What would settle it is specific and measurable: field trials that report pick rate per hour, fruit left behind, bruising and cull rates, and cost per acre across a full season, on more than one farm and more than one variety. Vendor demo footage does not answer those questions. Until that work exists and is published, this score leans on task structure and on the physical demands of the work. Our full approach is set out in the scoring methodology.
Good to know: a low evidence grade means uncertainty, not safety; it cuts both ways.
When the picture could change
Most likely after 2046 (8 in 10 of our scenarios). The replacement-year method explains what that window does and does not claim.
Two things could pull it earlier. First, crop breeding and trellis redesign: growers who plant uniform, machine-friendly varieties make the robot’s job far easier, and that shift is already underway in some orchards. Second, a sustained labor shortage with rising field wages, which shortens the payback period on a harvester and pushes growers to try one.
Two things hold it back. Gripper dexterity is the first: soft fruit and leafy crops punish any machine that squeezes or drops. The second is the economics of the seasonal calendar — a machine that works for six weeks a year has to earn its keep in six weeks, while repairs, dust, and heat eat into uptime. Small acreage and mixed plantings make that math worse, which is why mixed vegetable and nursery operations are likely to stay crew-based longest.
How to stay needed in crop and greenhouse work
Lean into the tasks machines leave alone. Hand-harvesting and handling fragile crops is the clearest one. Setting, checking, and fixing irrigation lines is the second, because water is where a season is won or lost. Propagation and plant care in a nursery or greenhouse — potting, grafting, pruning, spotting disease early — is the third, and it pays better over time.
Two skills raise your floor. One is running and maintaining machinery, from tractors and sprayers to the sensors and drones a farm already owns; crews that keep equipment running get the steady hours. The other is crew leadership: scheduling, training new pickers, and holding quality standards is work no camera does for you.
If you want to move sideways, the closest jobs are agricultural equipment operators, graders and sorters of agricultural products, and farmworkers with farm and ranch animals. The agricultural workers family page shows how the group scores together, and you can set any two of these jobs against each other on the compare tool. For a wider view of hands-on work, our list of jobs that mostly need a person is a useful next stop.