The Institution of Agricultural Engineers (IAgrE) artificial intelligence (AI) conference heard there was a fundamental difference in AI adoption between livestock and arable farming.
IAgRE member Declan Flynn, who specialises in AI and robotics for use in AI, told the conference dairy AI systems can deliver hourly savings, while crop-based AI applications often require seasons to demonstrate value.
READ MORE:听Government announces 'biggest animal welfare reforms in a generation'鈥�
The immediate benefit is evident in systems like DeLaval's automated milking technology, where AI adjusts parameters 20 times per second during harvesting operations.
Kieran Fitzgerald, vice-president, digital services at DeLaval, described the approach as creating a 鈥楩itbit for a cow', using AI to build digital twins of farms.
The company's disease risk models can identify conditions like mastitis and ketosis earlier than traditional methods, with AI penetration reaching 20-25% in mature dairy markets.
In contrast, precision agriculture applications such as variable rate fertiliser maps may show savings over two years, with spray applications delivering benefits twice per season.
Jonathan Henry, managing director of Garford Farm Machinery听demonstrated how Garford's Robocrop AI system uses artificial intelligence to combine colour, infrared, and depth information for precise weed identification, but the return on investment timeline differs significantly from livestock applications.
Paul Smith, a farmer using AI systems, reported simplified routines and measurable time and money savings, though the conference acknowledged that proving business cases remained challenging across different farming sectors.
Should farmers be paid for their agricultural data?
One of the most contentious discussions at the conference centred on who should benefit financially from farm data - and whether farmers should charge for it.
Professor Simon Pearson, Lincoln University, who holds the IAgrE Award of Merit听made a strong case that farmers should own their data and receive value in return.
He proposed establishing data trusts, structured as Community Interest Companies owned by farmers, which could then sell anonymised data to users while keeping control in farmers' hands.
However,听Fernando Auat Cheein, an international expert on robotics from Harper Adams University, offered a contrasting view, stating he did not see a viable business case for selling individual farm data.
He argued that data should be shareable, with the value depending on crop type and farm size.
The debate reflects broader concerns about data security as farms become increasingly connected.
Kieran Fitzgerald, vice-president, digital services at DeLaval听highlighted the responsibility companies face in protecting farm data and complying with regulations like the EU Data Act.
Morten Bilde, managing director at AGCO, stressed the importance of edge computing in harsh agricultural environments to mitigate such risks, suggesting that business-critical information should be available locally, not solely in the cloud.
Delegates heard that while standardisation and data sharing could accelerate AI development, the question of compensation and control over farm data remained unresolved.
READ MORE: Baroness Minette Batters' Farm Profitability Review: Industry reacts as review published
听
