Questions farmers should ask before bringing AI systems onto the farm
A representative of the Italian association of AI professionals urges farmers to establish how new AI systems work, who owns the data and who is responsible before buying them.
Artificial intelligence (AI) is already used on many farms, from smart cow collars and locomotion monitoring to laser weeders and automated slurry scrapers. Farmers considering such systems would do well to settle questions of privacy, data ownership and responsibility first. That is the view of Rafael Patron, president of the scientific-technical committee at the Italian Association of Artificial Intelligence Professionals (AIPIA).
AIPIA is an independent non-profit based in Rome. It brings together developers, data scientists, researchers and consultants, and keeps a public directory of AI professionals. Patron notes that farmers rarely hear an opinion from someone with no stake in the sale.
What the system can and cannot do
Patron gives the example of systems that use camera footage of housed cows to build an individual profile of each animal. In his view, such a system does not see a cow at grass in July, does not examine a foot, and cannot explain why a cow's mobility has dropped. Farmers should therefore ask the supplier which conditions the system is and is not designed to detect, what happens to the score when cows are out for the summer, and how many false alarms it produces. A system that flags too many cows gets ignored, while one that flags too few gives a false sense of security, which he considers worse.
Who owns the data
Another point that may lack transparency is ownership of the information collected. Daily footage and scores add up over a year to a detailed record of how a farm is run: stocking, bedding, feeding routines and how quickly problems are spotted. If a system is part of a buyer's scheme, farmers should find out whether the buyer sees raw scores, a summary or only what the farm chooses to share. They should also ask where footage is stored, for how long, whether it trains the supplier's algorithm, and what they can take with them when leaving. Patron stresses that answers should be obtained in writing.
Who makes the decisions
Before installation, farmers should agree with their vet which alerts trigger a visual check and which go straight into the herd health plan, and decide who checks the dashboard daily. Patron believes those who benefit most will be farmers who already know their cows well. He also suggests visiting a farm that has run a similar system for a full year.
