AI in agriculture: Advancing crops, insect protection and climate change forecasts

24 May 2024 --- Artificial Intelligence (AI) advocates say the technology can drive an “agricultural revolution” at a time when climate change and the rising human population weighs heavily on crops and F&B supply chains. Food Ingredients First explores the possibilities in the domain, the emerging innovations and their potential role in the agriculture value chain for fostering better crop development, crop monitoring and sensoring. 

One important challenge of tackling climate change is its impact on insects, on which nearly 75% of crops rely, says Omer Davidi, CEO of BeeHero, an Israeli company that leverages advanced data analytics, AI and low-cost IoT sensors to improve crop yields and increase profits for commercial crop growers.

“There's two aspects. First of all, it affects the bees, which are some sort of livestock and that's also affected by weather conditions. So having drier winters or having more extreme conditions puts some more stressors in bees and beekeepers to maintain strong and healthy colonies,” he tells us.

“The second aspect and we already see it in the fields, where we operate is the bloom time and the timeframe for pollination. So the more extreme weather conditions you get, you'll practically get a shorter timeframe for pollination activity.”

BeeHero's CEO Omer Davidi discussed the alarming decline of bee populations due to climate change and its impact on food production resilience.Some 40% of bee colonies collapse every year, and this affects honey prices, honey production and also agriculture, access to the right tools for beekeepers is important, he emphasizes

Fluctuations impact business
Ongoing climate change disruptions do require F&B business to adapt, Will Kletter, VP of Operations and Strategy at ClimateAi, agrees.

“Oftentimes when we think about climate risk, we think about tying that risk to declining yields. When we hear from our customers it is actually a slightly different story. We hear a story of volatility of a good year here and followed by a bad, wet year followed by a dry year, a significant uncertainty,” he said at the recently concluded F&A Next event at Wageningen University & Research in the Netherlands.

Fluctuating climatic conditions lead to issues like mycotoxins in wheat and other commonly consumed grains across the US, which he expects to steadily increase decade by decade. 

“This is caused largely by heat and humidity during the flowering period for crops so as the world warms up, we will see the active regions for this fungus continue to expand.”

US-based ClimateAi has built the “first climate resilience platform,” pioneering the application of AI to mitigate the impact of climate change. The platform forecasts the impact of climate change on supply chains, specifically food and agricultural supply chains, he explains.

He believes AI can help manufacturers combat such climate “inflexion points” as a “speed and effectiveness multiplier.”

“This could be accelerations and automation of different formulations or different breeding that you can simulate rather than having to go through trials. We work with a lot of seed companies, especially here in the Netherlands in the vegetable seed industry.”

BeeHero supports beekeepers with traceability and validating honey sources to address retailer concerns. “One such client has used one of our tools to identify future seed production regions based on the areas that they are successful today. We are able to reduce their validation time by one to two years, saving money and bringing speed to market more quickly.”

AI for better pollination
Meanwhile, BeeHero uses AI and IoT to protect bee colonies for better crop production through its IoT concept, which can go into the beehives to give beekeepers insights they didn't have before, underscores Davidi.

“An experienced beekeeper investigates the hives and then determines the problem. The AI component of BeeHero is pulling data constantly from the hive and ideally to replace this inspection process.”

The technology also helps the company compare the pollination activity of crops in certain geographies to indicate their potential impact on the F&B industry and crop yields.

“We collect data from inside the hive and also from the field, we can also count the number of bees that go in and out of the hive. We can actually model the efficiency level of every single bee in a colony and we're talking about 10s of 1000s of bees per hive.”

The company then aggregates the proprietary data collected from inside the hive and from the field, together with partners. This helps to gather information about weather activity and model the expected behavior of a pollinator in a specific environment and time of the year. Tracking helps catch any inconsistencies in the beehives, he explains.

BeeHero previously raised US$42 million to accelerate the growth and adoption of its “Precision Pollination Platform” and later unveiled its Pollination Insight Platform as an in-field sensing solution that measures pollinator activity in crops to improve pollination and increase yields.

Advancing precision agriculture
PhenoInspect, a German company that analyzes crops and weeds to improve crop production, uses AI and drones for “high-resolution agronomic scouting in agriculture.”

Germany-based PhenoInspect uses AI and drones for crop monitoring.“In breeding or in agricultural research, where a lot of crop monitoring is taken in the field to read new varieties to validate hypotheses, drones can be a very useful tool, Philipp Lottes, CEO of PhenoInspect said at the F&A Next event.

When the drone flies, AI can detect every crop (like wheat) on the field, put it into an executable application map, put it on a tractor and then execute in a targeted tractor that will spray insecticides “only where and when” they’re needed, he explains.

The company partners with federal inspectors to extract relevant information from the drone using AI and has developed a plant analyzer web interface which covers the entire data management. 

Preparing for safe AI adoption
The EU recently proposed an AI Act which is the first comprehensive regulation on AI by a major regulator anywhere.

A recent report released by Rabobank highlighted the AI’s potential role in transforming the food sector by refining operation and production, but also flagged risks to the industry, such as affordability issues, “AI washing” and the dangers of irresponsible use.

Food Ingredients First spoke to Katherine Dutmer, lawyer in Life Sciences and lecturer at the University of Amsterdam to understand how F&B companies can use it for enhanced efficiency while being safe.

Dutmer advises F&B companies using AI models, transparency is foremost when using Generative AI.“While AI can bring a lot of innovation, if you want to have useful, profitable innovation, you need to include it in the right way. Take care of the data,” she flags.

“The new AI rules are divided on a risk-based approach, for limited risk AI applications, for example transparency is foremost. For biometrics or other use of personal data considered ‘high-risk’, further risk mitigations and control measures are needed.”

The EU Commission can adopt new AI applications in high-risk categories in future. She advises companies to check “foreseeable misuse” of their AI models and have risk mitigation and control measures implemented.

“Train your people and document your AI developments.”

“Most companies are now dealing with the fact that they want to use Generative AI with considerable concern in a sensible way. It's best to use it in a contained environment,” she concludes.

By Insha Naureen, reporting from F&A Next, Wageningen University & Research, The Netherlands