- MIT receives up to $53 million to develop AI herbicides.
- The research aims to address the issue of resistant weeds and reduce reliance on chemical herbicides.
- Three teams will collaborate to screen and test candidate molecules and make improvements.
- The project also includes converting pesticides into biomass to reduce environmental impact.
A team led by MIT will receive up to $53 million from the U.S. Advanced Research Projects Agency for Health (ARPA-H) to develop AI-designed herbicides targeting resistant weeds, with a lower risk to humans. This work relies on generative artificial intelligence, which can learn patterns from existing data and propose new candidate molecules. The project is named "Precision Learning for On-target Weed control" (PLOW).
Christopher Voigt, head of the Department of Biological Engineering at MIT, is the principal investigator. He noted that glyphosate and atrazine face "a real crisis due to resistance mutations." According to MIT, over 1 billion pounds (approximately 454 million kilograms) of herbicides are used annually in the U.S., and certain weeds have developed resistance to the most common products. MIT states that exposure to these chemicals may be linked to cancer and metabolic and endocrine diseases, with small amounts of these chemicals entering food and drinking water.
Research Team and Collaboration
This funding is part of ARPA-H's Sustainable Technology for Efficient Agricultural Management (STREAM) program, launched by the U.S. Department of Health and Human Services on August 17, aimed at finding safer weed control tools. MIT's announcement indicated that three teams will share this work. A generative AI team led by Connor Coley, Regina Barzilay, and Alexander Tropsha will develop models to predict compounds or small interfering RNA (siRNA) that can disrupt the plant enzymes targeted by glyphosate and other herbicides.
Glyphosate inhibits the synthesis of several amino acids by blocking the EPSP synthase. These models will also consider synthesis costs, resistance risks, and potential harm to humans or ecosystems, with the expectation of generating billions of candidate molecules.
Environmental Considerations of the Project
The second group, led by Angela Koehler and Jaime Cheah, will screen these candidate molecules and provide feedback. Voigt stated that the screening data will be used for computational methods to facilitate improvements. MIT researchers have also released an independent AI molecular screening tool for drug discovery. The third group will test the screened candidate molecules in greenhouses, with conditions varying according to different farm scenarios. This group is led by Neal Stewart and Scott Lenaghan from the University of Tennessee, Patrick Tranel from the University of Illinois, Tim Willson from the University of North Carolina, and Andee Wallace from Robigoo.
The project also includes engineered bacteria and fungi aimed at converting pesticides into biomass to prevent them from entering food, soil, or waterways.
Additionally, a study from the University of British Columbia showed that fruit washes in laboratory tests could remove 86% to 94% of pesticide residues from apples. However, no results have been confirmed yet. MIT describes the project as a plan rather than tested molecules and has not provided performance data. This $53 million funding represents the maximum amount available. MIT also indicated that these alternatives could reduce U.S. dependence on imported phosphorus, a key component of glyphosate, but did not provide specific data. Initial greenhouse results for RNA-based candidate molecules are expected to be announced before the end of the first year.
Potential Impact of AI Herbicide Development
This research led by MIT aims to address the current issue of herbicide-resistant weeds in agriculture, showcasing innovative applications of artificial intelligence technology in the field. As resistance to traditional herbicides increases, developing new herbicides becomes crucial. This not only helps improve crop yields but may also reduce negative environmental impacts, particularly pollution of water sources and soil. Through multi-party collaboration, this project could provide new solutions for future agricultural sustainability and promote advancements in agricultural technology.

