Loading page…
The PLOW project will connect AI-generated candidates with laboratory screening and greenhouse studies. Initial RNA-based tests are expected by the end of its first year.
Listen to this story
MIT’s PLOW project, led by Christopher Voigt and backed by up to $53 million through ARPA-H’s STREAM program, is structured as an iterative discovery pipeline: generative models will propose chemical or siRNA candidates, lab teams will screen them, and results will feed back into model development. Researchers are optimizing for synthesis cost, resistance risk, and human and environmental safety—not just weed control—and also aim to reduce reliance on imported phosphorus used in glyphosate. These are research goals, not proven outcomes; initial greenhouse tests of RNAi candidates are expected by the end of the program’s first year.
The models may generate billions of candidates targeting plant enzymes used by existing herbicides, including EPSP synthase.
A separate team will test shortlisted molecules under different agricultural conditions after lab screening.
PLOW also plans living microbial systems to convert agrochemicals and formulation ingredients into biomass before they reach food, soil, or waterways.
An MIT-led team has received an award of up to $53 million from the U.S. Advanced Research Projects Agency for Health to pursue alternatives to today’s weed killers. Described by MIT News on November 10, the project will develop generative AI tools to search for candidates designed around weed control, human health and environmental safety.
The effort starts with a problem farmers already face: some weeds have acquired genetic mutations that make common herbicides less effective. More than 1 billion pounds of herbicide are used in the United States each year. These chemicals prevent crop losses and enable regenerative farming practices that protect soil, but resistance threatens their usefulness.
Health concerns add another reason to look for alternatives. MIT’s account notes that herbicide exposure may contribute to cancer, metabolic and endocrine disorders, and other adverse outcomes. Farm workers face exposure through frequent use; other residents can encounter small amounts through food and drinking water.
Christopher Voigt, head of MIT’s Department of Biological Engineering, leads the project, called Precision Learning for On-target Weed control, or PLOW. The award comes through ARPA-H’s STREAM program. Beyond health and resistance, the team aims to reduce dependence on imported phosphorus, an ingredient in the widely used herbicide glyphosate.
The planned models will propose chemical compounds or small interfering RNA, known as siRNA, that disrupt plant enzymes targeted by existing herbicides. Glyphosate, for example, interferes with EPSP synthase, an enzyme plants need to produce several critical amino acids. The researchers want new candidates that act on these targets while reducing the risk of resistance.
Killing weeds is not the only design goal. The models are intended to optimize for low synthesis cost, low risk of resistance, and low risk of harm to humans or ecosystems. Those are research objectives, not demonstrated properties of a finished herbicide.
MIT researchers Connor Coley and Regina Barzilay, alongside UNC-Chapel Hill’s Alexander Tropsha, will lead the AI group. The models are expected to generate billions of possible candidates. A second team, led by MIT’s Angela Koehler and the Broad Institute’s Jaime Cheah, will screen candidates and return experimental data to improve the computational approach.
Once screening narrows the candidate pools, greenhouse studies will examine how the molecules perform under different agricultural conditions. That work will be led by researchers from the University of Tennessee at Knoxville, the University of Illinois at Urbana-Champaign, UNC-Chapel Hill and Robigo. MIT expects initial greenhouse tests of RNA-interference, or RNAi, candidates to be complete by the end of the program’s first year.
PLOW also includes a parallel effort to reduce exposure to agricultural chemicals already in use. Researchers will develop living microbial systems intended to degrade agrochemicals before they reach food, soil or waterways. The proposed systems would convert chemicals and formulation ingredients into biomass, with the goal of reducing environmental and human exposure.
Loading discussion...
Join the conversation
Explain whether combining those goals strengthens the research or spreads it too thin.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.