How TED talks frame artificial intelligence
A TED talk on artificial intelligence tends to begin at altitude. The camera pulls back, the stakes scale up — transformation, disruption, the re-ordering of industries — and by the end the audience is somewhere between inspired and pleasantly alarmed. That register is genuinely useful: it makes complex ideas legible and gives people permission to think ambitiously. TEDx Grantham, operating inside the TED ecosystem, shares that instinct for ideas worth spreading. But this corner of Lincolnshire adds a complicating fact.
Right now, in polytunnels across the county, autonomous robots are scanning every row of tabletop strawberries through the night, counting roughly three million flowers and fruits per robot per shift — not as a pilot demonstration, but as a working part of commercial fruit production. The AI questions this raises are specific and consequential: how do you keep a navigating robot safe around seasonal workers? How do you verify a yield forecast that a farm's entire season depends on? How do you replace labour that accounts for half of production cost when your ability to raise prices is close to zero?
Those are the questions that rarely make it onto a TED stage. They arrive not as futures but as engineering requirements, safety standards, and labour economics — and they are being answered, field by field, in a county that grows 30% of the UK's vegetables and employs a food chain workforce twice the national average. What does AI actually look like when it enters food production in a place whose economy depends on it?
Why Lincolnshire's agricultural identity makes this matter nationally
The county's scale in English food production is not rhetorical shorthand. Greater Lincolnshire grows 30% of the UK's vegetables and 18% of its poultry, and holds more Grade 1 agricultural land than any other Local Enterprise Partnership area in England. In 2023, its agricultural output reached £1.1 billion — 10% of the UK total. The food chain employs 24% of the county's workforce, against a national average of 13%, and generates 21% of its economic output compared with 7% nationally. Those are not figures that belong in a regional boosterism leaflet; they describe a structural dependency. When this concentration of food supply originates in one county, the decisions made in its fields — which technologies to adopt, how quickly, and at what risk — carry consequences that reach well beyond Lincolnshire's borders. An AI deployment that works or fails here does not affect a peripheral corner of British farming; it touches the systems that underpin UK food security. That is the difference between Lincolnshire as a convenient local example and Lincolnshire as the place where the question is most directly live.
What the robots in Lincolnshire's fields actually do
At Riseholme, a 200-hectare working farm campus north of Lincoln, sits the institution formally recognised by the UK Innovation Strategy in July 2021 as 'the world's first global centre of excellence in agricultural robotics'. Lincoln Agri-Robotics (LAR), directed by Professor Simon Pearson and funded by UKRI, is not a research park adjacent to farming — it is embedded in it, which distinguishes it from a university robotics lab with distant ambitions.
The MeSAPro project (2020–2022), a collaboration between the University of Lincoln and Saga Robotics, exemplifies that embeddedness. Autonomous Thorvald robot fleets were deployed across real soft-fruit farm environments — not simulated ones — covering four distinct scenarios: monitoring, forecasting, harvesting, and transport. The safety engineering that accompanied this was formal rather than precautionary. Researchers conducted hazard analyses across all four scenarios, formally verified a sense-understand-decide-act pipeline, and integrated a Human-Aware Navigation system directly into Thorvald's commercial software stack. That last element was designed for a specific real-world condition: seasonal workers and robots sharing the same polytunnel space, in a sector where agricultural workers represent 1% of the UK workforce yet account for 20% of all workplace fatalities.
Thorvald itself has moved well past prototype status. The platform now treats approximately 30% of UK tabletop strawberries with UV-C light to control powdery mildew — a fungal disease that can devastate a season. At LM Porter, one UK farm, that treatment contributed to a documented 15% rise in production.
Behind the current deployments sits a longer-term talent pipeline. AgriFoRwArdS — designated the world's first EPSRC Centre for Doctoral Training in Agri-Food Robotics, run jointly by Lincoln, Cambridge, and UEA — signals a sustained institutional commitment rather than a single project cycle. The expertise being built here is not incidental to the region; it is being built because the region requires it.
The labour economics behind the technology
The numbers that explain why this is happening now have little to do with enthusiasm for technology. The UK soft fruit market is worth more than £1.3 billion at retail (Kantar), yet roughly 50% of total production cost is labour — the single largest input by some margin. Some 32,000 seasonal, typically migrant, pickers tend around 160,000 tonnes of fruit each year, and those workers are measurably harder to recruit than they were a decade ago.
Brexit has structurally reduced the pool of available seasonal labour from EU member states. That is not a contested interpretation — it is documented in industry cost models. The difficulty is compounded by weak pricing power: soft fruit growers have limited ability to pass cost increases on to major retailers, who set the terms. Labour costs rising, labour supply tightening, and margins already thin: that is the context in which UKRI backed the First Fleet project.
Led by Saga Robotics with Berry Gardens, the University of Lincoln, Oxford, and NIAB, the First Fleet explicitly targets the world's first fully autonomous, multi-modal seasonal fleet — not a solution to one task but a system designed to cover the full harvest-season repertoire. The ambition is framed as a structural response to a fragile model, not as automation pursued for its own sake.
What remains open is who benefits when the technology matures. Agri-OpenCore, the Lincoln-led initiative for an open-ecosystem harvesting platform, is presented as a route to lower development costs and wider access for smaller growers — but whether it delivers on that is still an economic and institutional question as much as an engineering one.
The human questions the technology is raising
The displacement question gets a genuinely complicated answer in Lincolnshire's soft-fruit sector. Seasonal picking work has contracted not because robots arrived but because the workforce it depended upon — largely drawn from EU member states — became harder to sustain structurally after 2016. Autonomous systems entering that context are not straightforwardly taking jobs from people who hold them; in significant part, they are filling roles that growers are already failing to recruit.
That does not dissolve the concern — it redirects it. The sharper question is whether the gains from autonomous deployment concentrate among operations large enough to bear the capital cost. A grower able to contract Thorvald fleets at scale is not in the same position as a smaller operation managing a few polytunnels against thin margins. Agri-OpenCore, Lincoln's initiative for an open-ecosystem harvesting platform, is an explicit attempt to lower that barrier — though whether it succeeds is still an economic and institutional test, not an engineering one already passed.
A separate case runs in automation's favour, rarely made publicly. Farm workers bear a share of workplace fatalities wholly out of proportion to their numbers — the injury statistics that drove MeSAPro's formal hazard analysis were not background data; they described an occupational reality that the Human-Aware Navigation system was built to address, with co-worker safety as a design constraint rather than an afterthought.
The hardest question is community. AgriFoRwArdS is building a postgraduate cohort at Lincoln, Cambridge, and UEA — engineers, researchers, and system designers whose roles differ entirely from seasonal harvest work in training, pay, and conditions. Those positions are real. Whether they will reach the communities that have long supplied the county's picking workforce is something Greater Lincolnshire has not yet answered.
What this means for readers across the region
Somewhere between a TED stage and a Riseholme polytunnel lies a gap that most AI commentary does not bridge. The ideas explored in conference centres — transformation, the shape of a future workforce, the ethics of automation — are genuine ideas. But the decisions that will determine how AI enters Lincolnshire's food system are not being made by keynote speakers. They are being made in UKRI funding terms, in procurement contracts between growers and Saga Robotics, in the safety certification frameworks that govern what a Thorvald fleet is permitted to do unsupervised among farm workers, and in what transition obligations — if any — will attach to the First Fleet when it moves beyond pilot scale into full commercial operation.
TEDx Grantham sits inside the county where those decisions are taking shape. The questions worth carrying away from this are not abstract ones about AI and humanity in general. They are specific: who sets the safety standards for autonomous systems operating in Lincolnshire fields? Does Agri-OpenCore's open-platform promise translate into genuinely wider access for smaller growers, or does it remain an aspiration written into a funding bid? And what will the First Fleet's first full commercial growing season actually tell the industry about whether a transition away from seasonal labour is economically irreversible once it begins? Those are the questions TED talks do not ask — and they are being answered here.
- [1] Agricultural robot – Wikipedia. https://en.wikipedia.org/?curid=11005995 https://en.wikipedia.org/?curid=11005995