Crop protection companies utilize AI to develop products for resistant weeds

By Patrick Thomas, Wall Street Journal

Farmers are aiming to add a high-tech tool in their never-ending battle against weeds: artificial intelligence.

Farmers have been losing ground in a decadeslong fight against intruding plants that have developed resistance to many go-to chemical sprays. The world’s biggest agriculture companies, including Bayer, Corteva, BASF and Syngenta, say the clock is ticking on efforts to develop new chemicals to keep the weeds and other pests, such as fungi and insects, at bay.

Some weeds can now withstand four or five different chemicals, says Bob Reiter, head of research and development for Bayer’s agriculture business. Farmers are turning to mixology to find the most potent weedkiller formula, and pesticide makers want to speed up what can be a yearslong process of identifying new herbicides.

″We’re getting a little bit into a point of desperation,″ Reiter says. ″We understand that we’re running out of things that are going to work.″

Bayer and rivals like Corteva and Syngenta say new AI systems are helping speed up the lengthy, complicated and costly process of bringing new chemicals to market. Their efforts include not only new herbicides for weeds but novel fungicides and insecticides as well. Syngenta estimates that AI will reduce the average time from discovery to commercialization by a third—to 10 years from 15 years—and likely decrease the number of lab and field tests by 30%.

Bayer is using an AI system, internally named ″CropKey,″ to help match the protein structure of a weed with a chemical molecule that targets that structure by sifting through data faster than humans. Molecules selected with CropKey’s help could have a higher success rate during field testing than conventional research, Reiter says. It gives the company an edge—like counting cards in a game of blackjack—and is similar to how drug companies are using AI to speed up the search for molecules that target a particular disease.

Companies say an advantage of molecules selected with AI is that they can be screened during the process for toxicity to humans—a critical point for pesticides sprayed on crops people will eat—as well as environmental safety and cost.

The system helped Bayer develop a new weed-killing chemical, called Icafolin, set to be released in Brazil in 2028. The company says it will be the first new herbicide in more than 30 years. The company’s early-phase research pipeline is also filling up with new ways to attack weeds, with three times as many such projects as there were a decade ago, it says.

Monsanto, now owned by Bayer, revolutionized the business of wiping out weeds in the 1990s when it began selling soybean seeds that were genetically modified to withstand glyphosate, a herbicide that kills weeds by halting their internal protein production. Use of glyphosate sprays, including Monsanto’s Roundup, soared.

But then weeds evolved, and glyphosate’s waning effectiveness left farmers increasingly reliant on other herbicides to fight superweeds, or plants resistant to more than one type of chemical. Roundup has separately cost Bayer billions of dollars from lawsuits alleging it causes cancer, which the company denies and maintains it is safe to use.

Sean Elliot, a sixth-generation corn and soybean farmer in Iroquois County, Ill., first noticed invasive waterhemp weeds on his farm in the early 2000s. Back then, Monsanto’s Roundup weedkiller curbed them with ease.

Two decades later, glyphosate can’t stop the weed, and Elliott says he is now worried that the waterhemp is overcoming another chemical he uses, 2,4-D. He mixes 2,4-D with a third chemical, glufosinate, to keep the waterhemp under control. In a few years, it might not be enough.

″It’s amazing how quick it adapts,″ Elliott says. ″If we don’t come up with something, and can’t control it, it’s such an invasive weed that you could be looking at major yield losses.″

Pesticide development can be more complicated than discovering medical drugs, says Bill Anderson, chief executive of Bayer, the German-based pharmaceutical and agriculture giant. It has been decades since a novel herbicide product was introduced, with companies largely focused on incremental improvements to the main chemicals already on the market.

″You need to be able to kill one kind of plant without killing other kinds of plants, but also without harming fish, insects, birds. It needs to break down in the atmosphere,″ Anderson says. ″The chance of getting it right without the computer methods is vanishingly small.″

Jay Feldman, executive director of Beyond Pesticides, a Washington, D.C.-based group that advocates using fewer agricultural chemicals, warns that spraying new chemicals onto weeds that are rapidly developing resistance to multiple herbicides could only exacerbate the issue—and create even stronger superweeds.

Agriculture companies’ approach will make older herbicides obsolete unless they are bundled with newer chemicals, he says, making farmers solely reliant on whatever new seed and chemical product a company is offering. ″This has created a pesticide treadmill,″ Feldman says.

As recently as five years ago, companies could spend a year screening hundreds of thousands of chemical compounds. Potential compounds would be tested in labor-intensive processes in laboratories and greenhouses to see how they interact with other plants, animals, humans and the environment, and if they even work on the targeted pest, says Shaun Selness, head of emerging technologies for Bayer’s agriculture business.

″You spent two or three years in the field scaling up, doing a lot of regulatory studies, come all that way only to find out we really can’t bring that forward,″ Selness says. ″It happened pretty frequently.″

Leaning on AI to help analyze and screen chemical molecules can help shrink the process down to about two to three months and predict possible toxicity problems earlier in the development process, he says.

Syngenta, the largest pesticide seller in the U.S., says it is taking a similar approach to new herbicides and insecticides—it uses machine-learning models for all its research projects to find new active ingredients.

The technology helps the company better evaluate not only the new products’ effect on the environment but also whether they can be produced at a lower cost, says Camilla Corsi, head of Syngenta’s crop-protection research. ″It helps to address all of the challenges our industry has when it comes to chemical innovation.″

Elliott, the Illinois farmer, welcomes any new technology that he thinks could help save his crop in future years.

″It’s a must,″ he says.