MIT-Led Team Awarded $53 Million to Develop AI-Driven Alternatives to Traditional Herbicides

A multidisciplinary team led by MIT has secured up to $53 million from the U.S. Advanced Research Projects Agency for Health (ARPA-H) to develop next-generation weed control solutions. This initiative, called Precision Learning for On-target Weed control (PLOW), focuses on designing herbicide alternatives that address the growing challenge of weed resistance and potential health risks associated with current agrochemicals.

Herbicides such as glyphosate, used widely in the United States with over one billion pounds applied annually, function by disrupting crucial plant enzymes like EPSP synthase. However, widespread use has led to genetic mutations in weeds that confer resistance, threatening crop yields and economic stability in agriculture.

The PLOW project leverages generative artificial intelligence to model and predict new molecules—either chemical compounds or small interfering RNA—that inhibit the same enzymatic targets as existing herbicides but with improved specificity and safety profiles. These AI-generated candidates will be evaluated for effectiveness, ease of synthesis, low resistance development potential, and minimal human and environmental toxicity.

The research consortium includes experts from MIT, the Broad Institute of MIT and Harvard, University of North Carolina, University of Tennessee, University of Illinois, and agricultural biotech company Robigo. The program is structured into three segments: AI model development, high-throughput molecular screening, and greenhouse trials to validate herbicide effectiveness under real agricultural conditions.

Complementing the herbicide discovery efforts, the team is engineering microbial systems capable of biodegrading agrochemicals before they reach soil, water, or food substrates. Such biological remediation could reduce human exposure to harmful residues and enhance environmental safety.

The project’s first phase anticipates completing greenhouse testing of RNA interference (RNAi)-based candidates within the first year. Despite promising approaches, challenges remain in validating long-term ecological impact, scaling production, and ensuring regulatory compliance for novel bioengineered weed killers.

Through this innovative convergence of AI, synthetic biology, and chemistry, the PLOW initiative aims to create sustainable weed management solutions that support regenerative farming practices, mitigate health risks, lessen dependency on critical raw materials like phosphorus, and maintain high agricultural productivity.

Funding for this research is provided by ARPA-H under agreement number 140D042690030. The viewpoints presented do not represent official U.S. government policy.