The digital farmer: How AI enabled agriculture will help feed the world

Efficiency gains need to be made in the agriculture sector if it is to feed the two billion people that will be on the planet by 2050.

Farming is often a delicate balancing act of doing something at exactly the right time; get it wrong and an entire crop can be ruined.

This is why the industry needs innovation to improve processes and output and to make more accurate predictions and decisions.

This is where AI can help. According to a report by Mind Tree, agriculture is seeing rapid adoption of AI and machine learning for both agricultural products and in-field farming techniques, with cognitive computing to become the most disruptive technology in the sector.

By harnessing different data from remote sensors, weather, satellite information, LiDAR, and combining it with industrial AI farmers can derive new actionable insights before problems arise.

For example, in the Netherlands, Connecterra’s Intelligent Dairy Farmer’s Assistant collects data from sensors located on a collar worn by cattle.

#Dutch technology is helping to make U.S. #farming more efficient by combining #artificialintelligence and motion sensors to monitor cows, after having piloted the program in Europe for years https://t.co/5FKfmRAOvy pic.twitter.com/30s90vEszP

— NFIA North America (@NFIA_USCanada) April 26, 2018

Feeding the data into Google’s AI software, TensorFlow, the system can then detect patterns and provide insights to farmers about when there is a problem and action is needed.

For example, through monitoring a cow’s behavior, it can determine if it is sick and alert the farmer. Such remote monitoring saves farmers significant amounts of time as normally they would have to check the temperature of every single cow manually to monitor their health.

According to Intel, AI technology is being developed and researched at NatureFresh Farms, a 20-year-old company growing vegetables on 185 acres between Ontario and Ohio, that can tell by a picture precisely how long a yellow seedling of a tomato flower will take to grow.

This information helps the sales team know exactly how many tomatoes will be available for sale at certain dates, knowledge that can directly benefit the bottom line.

Furthermore, Israel-based Prospera says it can provide sales and purchasing teams with 95% accurate yield predictions, maximizing farm output through optimal variety allocation, planting and harvest planning, inputs selection, and protocol enhancement. According to the firm it analyses $5bn worth of produce annually through 50 million data points.

Other major companies to enter the market include IBM, which has just inked a deal with The National Institution for Transforming India to develop a crop yield prediction models using artificial intelligence.

NITI Aayog partners with @IBM to develop crop yield #prediction model using #AI The idea is to use AI to improve efficiency in resource-use, #crop #yields and promote scientific #farming https://t.co/GdWPLyHQAZ

— Matthew Hill (@kiwi_matt_nz) May 7, 2018

The use cases for AI are far and wide. Potentially algorithms can also help identify plant disease, detect pest infestations and help automate farm equipment.

Considering time is an all-important factor in farming, taking out some of the guesswork is extremely beneficial. However, AI does have its downsides. As Arka Bagchi, Associate Consultant, Mindtree, notes in his report the cost can be prohibitive and there is a need for it to come down to enable mass adoption or for an open source platform to be made available.

Furthermore, it requires a lot of processing power and upfront investment. But given the scale, farmers often work at and the consequences of getting it wrong, it is an investment that many will deem worth making.