# Agriculture & Food Systems: Weekly AI Update

## Smart Water Management Changing Farming

Farmers around the world are learning to use artificial intelligence to save water. At a big farming conference in Dubai in December, experts talked about how AI is helping solve one of farming's biggest challenges. Right now, farms use more than 70 percent of all the fresh water people use, and a lot of it gets wasted. Faissal Sehbaoui, who runs a farm technology company called AgriEdge, says AI can change this. Instead of farmers guessing when to water their crops, AI can look at soil sensors and pictures from space to know exactly when plants need water. This is called moving from reactive to predictive water management, which means planning ahead instead of fixing problems after they happen.

The technology works by combining many different types of information. AI can read data from underground sensors that measure how wet the soil is, satellite photos that show the whole field, weather information, and lots of other clues. Then the AI tells farmers what to do next. "Thanks to AI, we can make the best decisions," said Sehbaoui. "We can adapt our irrigation programmes by taking into account all the data we combined and intervene before the plant experiences water stress."

## Helping Farmers in Poorer Countries

One big problem with AI farm technology is that it can be very expensive. Putting sensors all over a small farm costs a lot of money that many farmers, especially in Africa, don't have. However, scientists are finding solutions. Instead of expensive sensors, they can use free pictures from satellites in space. AgriEdge created two different AI tools: one with sensors for rich farmers, and one using satellite pictures for smaller farmers. "We calibrate the satellite models using field data," explains Sehbaoui. "The precision is not the same as sensors, but it is far better than having no data."

Another company called SunCulture is using something called carbon markets to make farm technology cheaper. When farmers use tools that are good for the environment, they can earn money from carbon credits. This money helps pay for the technology. "Affordability is a huge issue but the carbon market has been really useful," said SunCulture's leader Samir Ibrahim. "We've been able to now give a commercial subsidy to our customers to level the playing field."

## Scientists Create Tool for AI to Learn Better

At Berkeley Lab in California, scientists made something called EcoFABs, which are small plastic boxes about the size of a takeout container. These boxes let scientists grow plants and study the tiny organisms that live in their roots in a way that computers can understand. The exciting part is that scientists in four different countries—Australia, Germany, and the United States—all used these boxes and got the same results. This is super important for AI because AI needs to learn from information that is clean and consistent.

One scientist named Vlastimil Novak explained why this matters: "If you want to make meaningful predictions about microbes and plants, especially with future AI models, you need clean, consistent datasets. EcoFABs provide exactly that." The boxes help scientists understand how helpful microbes can make crops healthier and help soil stay rich. This knowledge could help farmers grow food with less chemical fertilizer and grow better crops in poor soil.

## Robots Still Learning to Farm

While AI is getting smarter at many farm jobs, researchers in Canada say that self-driving farm machines still have a lot to learn. A scientist named Jay Wang at the University of Manitoba is studying how robots can work in farm fields better. Right now, self-driving cars on roads work pretty well, but farm fields are much harder, he explains.

The problem is that farms are bumpy, muddy, and always changing. When it rains, the ground gets softer. Leftover crop pieces can be in the way. Different parts of the field are packed down differently. A road is flat, has painted lines, and everything stays the same. A farm is the opposite. Wang and his team are using AI called Gaussian process to help machines understand and adapt to these changes. The robots have special sensors like GPS and LiDAR that help them know where they are. As they move through fields, they collect information about how the ground acts differently than expected. Then AI uses this information to get smarter.

## AI Needs to Feel Helpful, Not Fancy

Experts say the most successful AI in farming won't be the fanciest or most complicated. Instead, it will be AI that farmers trust and that actually helps them make more money. Hunter Swisher, who runs a company called Phospholutions that works on soil health, says AI should help farmers make decisions based on real results, not just guesses. He explained that for a long time, farmers made decisions based on tradition or what they thought was safe.

"AI creates the opportunity to make decisions based on verified performance across time and place," said Swisher. However, he also warned that AI shouldn't make farm decisions more complicated or confusing. "The most impactful AI will not feel revolutionary. It will feel dependable," he said. This means AI should work quietly in the background, helping farmers every day with real problems like using fertilizer more efficiently or protecting crops from disease.

Experts also say that real progress in farm technology happens when everyone works together—farmers, technology companies, stores that sell farm supplies, and government leaders. They need to stop arguing about ideas and start asking simple questions like "What actually works, and how do we use it?" When this happens, farms can produce more food using fewer resources, which helps both farmers and the environment.