Back to
Can artificial intelligence help to feed the world?

Can artificial intelligence help to feed the world?

Article by Article by Dr Michael Snowden, CEO, OneNet Limited

Published on 23/08/2018


Can global food production double before 2050? - Writes Dr Michael Snowden

If not, the UN forecasts that famine from population growth and a deteriorating environment will likely have devastating effects as global population grows from 7.6 billion to 9.8 billion people.

Climate change and declining water availability are lowering productivity. Crop yields are plateauing and total global land under cultivation is decreasing. Societal attitudes towards food traceability and security add more pressure. Agricultural growth must now be sustainable. We cannot just plant more crops or breed more cattle.

What is the solution? Simply put, we need greater efficiency to produce "more with less". This is the idea behind "precision agriculture" which uses innovation to increase yields and decrease inputs, while minimising the environmental impact.

Precision agriculture uses satellites, drones, GPS, Internet of Things (IoT) sensor devices, robotics and automation. The growth of artificial intelligence (AI) in precision agriculture results from falling costs and rising capabilities. Increased sensor density, improved connectivity to the Internet, remote sensing via drones and satellites, together with cloud computing, all combine with AI to create high-productivity solutions.

A great example of AI in precision agriculture is the growing use of autonomous robots, or "agribots", which handle essential tasks such as harvesting crops at a higher volume and faster pace than humans can.

Benefits include constant work rates under a wide range of difficult environmental conditions and a reduction in the use of chemicals. An increasing scarcity of agricultural labourers and peak seasonal demand for human resources create further pressures, which may be reduced with agribots. Examples include robots that can precisely spray herbicides, pick strawberries and milk cows.

AI, used in crop and soil monitoring, processes data captured by drones and sensors to monitor water stress, nutrient conditions, plant population, soil moisture content, pests and diseases.

SkySquirrel, which uses drones in vineyards to help improve yields and lower costs, is a good example of the use of AI. Users pre-program the route and the drone uses computer vision to record images for analysis.

Data, captured by the drone, goes to a computer cloud for processing. AI software, or algorithms, analyse the images to report on the health of the vineyard. The company claims it can scan 30 hectares in 30 minutes and provide data analysis with 95% accuracy.

Predictive analytics using AI can help to predict disease outbreaks, the optimal time to plant seeds and improve supply chain efficiencies.

The firm FarmShots, for example, analyses agricultural data captured by drones and satellite images with the aim of detecting diseases, pests and poor plant nutrition on farms. The company claims it can inform users exactly where fertiliser is needed and reduce the amount of fertiliser used by nearly 40%.

Other applications of AI in agriculture include driverless tractors, which pass century-old machines over to robots to reduce human effort and work larger areas for longer hours.

Automated irrigation systems overcome the heavy reliance on historical weather systems to predict required resources. These systems use real-time machine learning, a form of AI, to maintain desired soil moisture conditions to increase yields.

Benefits include lower water and energy consumption. With 70% of the world's freshwater used for agriculture, the ability to better manage and reduce consumption will have a big impact.

As fewer inputs are required to gain higher outputs, almost everyone benefits. With less herbicide, pesticide and fertiliser used in agriculture, pollution and environmental degradation will be lower.

As farmers enjoy lower input costs, competitive forces will likely lower consumer prices, raise demand and still maintain farmers' profitability.

The main losers will be the vendors of inputs and low-wage agricultural workers.
The ultimate goal of AI in precision agriculture is to coordinate different precision agriculture systems so that recommendations are automatically produced which allow farmers to act immediately upon them.

AI is now turning precision agriculture into decision agriculture.

as seen in the Herald, 23 August 2018