Robots are already working on Lincolnshire farms
Inside a polytunnel on a Lincolnshire soft-fruit farm, a robot moves along the rows without a GPS signal to guide it — navigating by lidar and computer vision, locating fruit, logging data, carrying harvested strawberries back along the same path a human picker would walk. This is not a laboratory demonstration. As of May 2026, JABAS.AI — a spinout from the University of Lincoln's Centre for Autonomous Systems — is running exactly these trials with six commercial farm operators, its GPS-free fleet system designed specifically for the enclosed, signal-poor conditions of polytunnels and under-canopy growing.
JABAS.AI is one piece of a broader deployment already under way across the county. Platforms such as Dogtooth, Mi Machines, and Xihelm have placed strawberry-picking robots on Lincolnshire farms at scale — up to 20 machines documented on a single site. Lincolnshire grows roughly a quarter of the UK's vegetables and farms more than half of England's Grade 1 agricultural land, which means the choices being made here ripple through national food supply.
The question this article pursues is not whether farm robots are coming. They are here. The harder question is what this transition means for the tens of thousands of people whose seasonal labour has, until now, made the harvest possible.
What LIAT actually builds — and why it matters that it's physical
Picture a commercial mushroom shed: low light, high humidity, rows of substrate-packed shelves, and workers crouching to locate and cut individual Agaricus mushrooms by hand. LIAT — the Lincoln Institute for Agri-Food Technology at the University of Lincoln — is developing what it describes as the world's first commercial mushroom harvesting robot to work in exactly this environment. That single detail captures something the word 'agri-tech' routinely obscures: this is not a sensor measuring soil moisture or a dashboard forecasting yields. It is a machine doing the physical act a person used to do, in the dark and the damp.
The broader LIAT portfolio follows the same logic across multiple crops. The ARU (Agri-Robotics Unleashed) project integrates robotic harvesting arms directly into the architecture of soft-fruit greenhouses rather than retrofitting them. A fully automatic blueberry harvester has reached TRL 7 — near commercial-ready. ARWAC is an autonomous machine for open-field work; ISILDUR handles plant processing in forest nurseries. AGRI-OPENCORE goes further still: rather than producing a single product, it provides what LIAT calls the world's first open platform for other manufacturers to build their own robotic crop harvesters — a deliberate attempt to scale the whole sector.
Underpinning several of these projects is the GPS-free navigation research strand that gave rise to JABAS.AI: £4 million invested since 2017, with findings published in Nature. That investment history matters because it moves the technology beyond a press release into independently verified engineering.
The distinction between data tools and physical machines is the article's load-bearing point. Robots that carry, cut, and sort alter what human hands are needed to do — and that is where the labour question becomes concrete.
How Brexit reshaped the seasonal picking workforce
The numbers are blunt. Before Brexit, around 63,000 migrant workers arrived each year to pick British fruit and vegetables. Afterwards, that number was capped at 45,000, with each worker restricted to a six-month stay; by 2025, the horticulture visa quota had been trimmed further to 43,000. For farms already running on tight margins and tighter schedules, the arithmetic was punishing. In 2022, the shortfall was severe enough to leave an estimated £60 million of food to rot in the fields — a figure that gave the push for automation its political urgency and commercial logic.
The labour crisis is real. But it is worth pausing on what it frames out of view. The seasonal workforce was not conjured by a visa system and then removed by it. These were workers performing highly dexterous, physically demanding tasks — timing a pick by the give of a strawberry, navigating a row in heat, sustaining that judgement across a ten-hour shift. Simon Pearson, director of LIAT, has said publicly that robots cannot yet match the picking rate of skilled human pickers, and that society is wrong to call this work 'low skilled'. The labour crisis created demand for machines. It did not make the workforce that preceded it any less real.
Thirty-four jobs and a £3.8bn cluster
Here is the arithmetic that promoters of the cluster rarely lead with. The Ceres Agri-Tech pipeline — five spinouts from the Lincoln, Cambridge, and UEA partnership — has created 34 high-value jobs to date. The wider Greater Lincolnshire agri-food and agri-tech cluster carries a self-described £3.8 billion GVA and is billed as Europe's largest agri-food automation and robotics cluster. Both figures are true simultaneously.
That gap matters more in Lincolnshire than it would almost anywhere else in England. Food accounts for 24% of all jobs across the region — roughly double the national share of 13% — and 21% of its economic output against a national figure of 7%. When a sector this structurally dominant begins to mechanise, the workforce exposed to that transition is large relative to the county's total economy.
The spinout ecosystem is young, and projections are genuinely ambitious: the LINCAM initiative backed by UKRI points to hundreds of high-skilled technology roles over the coming decade, with a stated aim to double the cluster's economic contribution. Those targets may prove realistic. But they describe a future that has not arrived. In the present, the roles being created are engineering and software positions; the roles that automation targets are the tens of thousands of seasonal picking jobs on which the region's food supply currently depends. The scale of one does not yet approach the scale of the other.
The question the agri-tech story doesn't answer
The agri-tech story is told almost entirely in the voices of researchers, investors, growers, and government ministers. What is conspicuously absent from the public record — press releases, policy documents, industry profiles, and academic coverage alike — is any substantive account from seasonal pickers themselves: what the arrival of robots means for their income, their routes back, whether 'upskilling' is a real offer or a rhetorical one. No granular count of picking roles already lost to automation in Lincolnshire exists in that record either. The transition is documented from the technology outward; it is not documented from the worker inward.
That absence has a name in the academic literature. A 2025 paper, Scarce, suffering, saved, uses the term 'digital saviourism' to describe a recurring pattern in how farm automation is framed: workers appear in the narrative as problems to be solved — labour shortfalls, visa complications, cost variables — or as fortunate beneficiaries of technological progress. What they rarely appear as is agents with a stake in shaping the change. The study, based on interviews with agri-tech investors, entrepreneurs, and growers, argues that this framing is likely to reproduce racialised inequalities already embedded in the seasonal migration system.
The automation narrative tends to treat labour shortage as a supply problem: not enough pickers, so build robots. The structural conditions of that workforce — including exploitation concerns documented by the House of Commons Library — sit outside that frame. The question of what is actually happening to the people who do this work remains, in the public record, largely unasked.
Carrying fruit versus picking it — where the transition actually stands
The honest description of where automated harvesting sits in Lincolnshire today is not replacement but augmentation. JABAS.AI's current deployment moves harvested fruit through polytunnels; a human picker still makes the decision to select and detach each berry. The fully automatic blueberry harvester has reached TRL 7 — near-ready by engineering standards, but not yet operating at commercial volume. The direction of travel is clear and the investment behind it is sustained; the question is not whether this changes the work, but how fast and on whose terms.
That distinction matters because it locates the transition at a specific moment. There is still a worker in the polytunnel. The robot carries what she picks. The moment when that changes — when the judgement about ripeness, the reach, the detachment all transfer to the machine — is approaching, but it has not arrived uniformly or at scale. When it does, the tens of thousands of seasonal pickers working on Lincolnshire farms will have a material stake in how that transition is managed: retraining pathways, income continuity, honest accounting of which roles are genuinely augmented and which are progressively hollowed out. None of these appear in the cluster's public investment story. The robots have a roadmap. The pickers who currently fill the gap the robots are built to close do not.
- [1] Scarce, suffering, saved: farm automation as digital saviourism. (2025). https://doi.org/10.1332/27324176y2025d000000045 https://doi.org/10.1332/27324176y2025d000000045