How can artificial intelligence act as a helping hand for land managers?

In a recent report, we discover how artificial intelligence can offer significant productivity improvements to landowners who are willing to invest. By Sarah Gibbons
ai farming

With a wide range of industries adopting ChatGPT and related large language models, landowners and farmers are increasingly assessing the possible benefits of artificial intelligence (AI). The potential is clear - from sharper land valuations for sellers and data that enables enhanced biodiversity programmes for environmentalists, to better animal welfare and crop production.

The AI potential

A handful of early adopters are leading the way. Jason Beedell, Rural Research Director at estate consultants Strutt & Parker, says some landowners and famers fear AI may force them to “relinquish control over yet another aspect of their business”. But he and others are seeking to allay fears that using these systems means “farmers are being farmed for their data and don’t have ownership of it”. While AI has been mulled for years, Jason says landowners and farmers can now benefit from AI advice.

AI has arrived big time in terms of land management with so many examples. It’s really exciting and starts a very necessary conversation

Jason Beedell, Rural Research Director at Strutt & Parker

Indeed, AI brings “significant benefits”, reducing “low-grade human work”, and increasing accuracy and consistency within decision-making. AI algorithms can aid multi-criteria decision-making when it comes to complex issues such as land use strategy, absorbing all available data and proposing optimal land uses.

In estate and property valuation, AI can drive an automatic valuation model – gathering public information to provide data analysis on land prices for comparatively sized estates in the same region. “It’s exciting in terms of efficiency for property management,” Jason says.

A helping hand

AI can aid regulatory compliance for landowners and farmers, notably on conservation and biodiversity. Complex planning laws focusing on Biodiversity Net Gain can be handled more effectively using AI programs, according to one major vendor.

In an April 2023 blog, Steve Cooper, Product Developer of AI platform provider Informed Solutions, said AI and associated data science “can dramatically improve the speed and quality of planning decisions, helping us to protect, harness and responsibly develop land assets in smarter, greener and more economically viable ways”.

Ensuring environmental factors are woven into planning applications requires, he said, “the collation and analysis of huge amounts of data, often derived from disparate sources, whether [...] land use surveys, conservation data, or development case management information.”

Such data is typically not integrated, unstructured and in different formats, impeding decisions and proposals. A land block might be “a breeding ground for migratory birds [that] contains rare flora and fauna.” Planning applicants need solid data – planning officials will be using similar systems to make their decisions.

AI applications

Informed Solutions’ InformedDECISION system uses AI and natural language processing (as does ChatGPT) to offer land use guidance through integrating and leveraging mapping, data analytics and satellite imagery.

Satellite communications provider TS2 SPACE from Warsaw, Poland, says AI can help farmers analyse weather forecasts, soil conditions and other sources to determine when to plant, water and harvest crops, boosting health and yields.

Dr Tom August, a computational ecologist with the UK Centre for Ecology and Hydrology in Wallingford, Oxfordshire, has helped develop AI-driven platforms such as E-Surveyor and AMI-trap, which help users identify insect pests and invasive weeds by comparing images against a sample database.

“This sort of AI has come a long way in a short space of time,” he says. “We can now use these AI tools to help people with very limited experience identify plant and insect species on their land.” This includes species aiding natural pest control: “You can see which pollinators and natural enemies are being supported by your habitats, and how your habitat quality compares to national standards,” he adds.

Systems improve as they are fed more images of plants and insects. This gives landowners and farmers “verifiable data on change in habitat quality over time and evidence of where your land is performing well for nature, and where improvements could be made”.

Forestry assessment models using AI and drone imagery to measure above-ground tree biomass are “much more accurate than using models generally based on a few trees”, helping landowners claim carbon credits, notes Jason.

AI can help breeders, too: “By analysing large datasets on genetic information, phenotypic traits and environmental factors, AI algorithms can identify the most promising candidates for breeding programmes.” Jason says this can help to develop crops with higher yields and improved pest, disease and climate change resistance.

AI algorithms can also incorporate machine learning techniques and identify data patterns and relationships that humans might miss when it comes to weather forecasts, helping farmers and land managers better respond to weather conditions. Google’s DeepMind uses AI for ‘nowcasting’ – pinpointing weather events two hours ahead.

ai farming (2)

Adopting the tech

Despite the benefits, uptake is low. “There will always be early adopters - about 5% of landowners or rural business managers will take the risk then it filters down into the rest of the rural business population,” says Jason. For farms or estates, cost will be challenging where specialised AI is needed.

James Kavanagh, Land and Resources Director at the Royal Institute of Chartered Surveyors Standards & Practice, says AI and machine learning systems will help, but may not transform existing practices. “Users of AI in land and property will, more or less, find themselves doing things much as they always have but with an AI ‘adviser’ who may be able to find a solution based on previous examples across a much wider data sample, which users did not have access too previously.”

For those fearing a lack of human oversight, he adds: “AI will always need an expert human to check its output, we should think of it as an ‘apprentice’ requiring guidance.”