Rise of robocrops

The following feature first appeared in the most recent edition of Landworker – you can read the full issue online here.

Vast fields, monitored by drones and sensors, with all planting, crop management and harvesting carried out by robotic tractors controlled by AI and a few workers sat behind screens miles from the farm.

This all sounds like the plot of some rural science fiction story, but it is fast becoming a reality, and one that farmworkers need to be aware of.

Robotic systems using artificial intelligence (AI) have been trialled in fruit farms for years, and now the tech is moving out of the glasshouse and onto the fields.

Researchers have already developed AI systems to help reduce emissions associated with livestock farming and land use and help farmers tackle disease.

AI is being developed to improve so-called “precision farming” where machine learning algorithms and data analytics are used to give farmers and farmworkers more information to help crop management decision making around planting, irrigation, fertilisation, and harvesting. 

Data collected from field-based sensors can monitor soil moisture, temperature, and nutrient levels, helping AI systems recommend actions for the best results. This can cut down on waste while improving crop yields.

Once the best crop management plans are settled on, robot tractors can head into the fields to carry out the required work, monitored by an array of cameras and IT systems, with drones flying overhead to map the results.

Companies are developing ways to introduce AI into the cab. Dutch business Farmatronic claims that the “driverless tractor provides everyday farmers with more free time”.

John Deere is advertising a “Fully Autonomous Tractor” with the operator using an “Operations Centre Mobile” so they can monitor the tractor anytime, anywhere with a smartphone or tablet. This gives the user alerts if the tractor senses an obstacle or mechanical issue, with live video access for “real-time views of what the tractor is doing at any time.”

Firms are also offering “autonomous tractor retrofit kits” to upgrade existing machinery, reducing the costs of new equipment.

Peterborough company Garford Farm Machinery has been working on “Robocrop AI” which, the website says, is a “next generation, multi-modal vision guidance system” that uses AI to improve weeding between crop rows and has a “multi-modal camera”, “combined with an AI neural network, which is trained to identify and recognise the crop plant and stem location, then distinguish it from weeds and non-crop plants, using colour, as well as leaf shape and profile.”

There are concerns the adoption of AI in the countryside may see skilled farmworkers replaced by technology that can be used by anyone with a phone app, and that large, automated machinery working on land with public access could result in liability issues for farmers in the event of an accident.

Farmworkers will need to be trained to use the new technology, not a bad thing in itself, but it could be an issue with an ageing workforce needing to learn new skills around data analysis, interpretation and gaining an understanding on how these systems will work.

While some see this move encouraging younger people into the sector, there must be a just transition for existing farmworkers, with employers ensuring full training is given if they plan to increasingly use AI in the fields.

However tractor driver and Unite agriculture chair, Steve Leniec, feels there is no need to be too concerned yet.

Steve said, “I’ve not come across AI when it comes to managing crops. Yes, there have been a lot of developments in horticulture, with micro-robots in the glasshouses, but I don’t see a rapid uptake in the wider farming industry.”

Steve is aware of an increased use of robotics in farming, for example in milking sheds and GPS systems in tractors, but not AI.

“I feel the cost of labour, against buying and introducing the new technology, is not something that is practical for many farmers at the moment, though there will always
be some early adopters.”

Though not an issue in the fields yet, Unite is very clear that when changes do happen, workers must be fully involved. 

Responding to the recent AI Opportunities Action Plan, Unite warned that workers must not become the victims of the AI revolution, and must have a say in its future. If rushed AI can create new risks and leave workers feeling alienated and demotivated.

Unite general secretary Sharon Graham said, “The UK needs to embrace new technologies, but we also need proper protections from AI’s pitfalls and workers must have a say in how that happens.

“The introduction of AI in the workplace must be something that happens with workers and not to workers. Government, employers, and unions all need to be working together to avoid the potential dangers of workplace AI.”

Unite is particularly concerned about how AI programmes are increasingly being used for surveillance purposes on workers and has produced a guide and model agreement for officers and representatives. 

This says that “AI alongside other technological advances will only benefit the whole of society if linked to strong protections for both citizens and workers.”

The guidance highlights the need for new rights and protections, new regulatory controls and industrial standards and new agreements.

These would protect the public and workers in areas where AI could be used “in safety critical areas to replace physical checks and human monitoring of fail-safe systems” and agreements “covering areas from job security, shorter working time and job sharing without loss of pay”, to “rights to retraining, upskilling and lifelong learning.”

Artificial intelligence has the potential to improve the workplace for many people, including farmworkers, and no one is suggesting a ‘SkyNet-style’ control is just around the corner. But workers must be at the centre of any developments and not left behind in the rush towards a brave new world. 

See the Unite website for advice.

By Keith Hatch